MySQL A Threat to Bigwigs?

Disoculated writes "Is MySQL a threat to bigwigs? is the question asked in CNN's technology section. The article notes that MySQL is running perhaps 20% of the web databases but its revenue is merely 0.02%... yet the company is still making money and putting out an excellent product. Is this a sign that the database market is in for a drastic change? Of course, there's no mention of PostgreSQL or mSQL, but I guess that's typical."

Comments

Version 4 Will Tell

[Daveman692]

It all boils down to how MySQL 4 is looked upon. Big sites already use MySQL, as seen here, but will version 4 have enough features, be robust enough, and provide the support to convince those running things such as Oracle to switch.

Re:Version 4 Will Tell

[jeepthang]

Yip, you're 100% correct. The current version of MySQL simply can not compete with the Oracle's of the world. When MySQL launches version 4, they will have nice features such as sub-selects, and the undeniably important transactions. Only then will we see if MySQL can take on the big boys.

Re:Version 4 Will Tell

[letxa2000]

I think the issue is that MySQL is very adequate for 99% of all users including most large enterprises and certainly most websites. Even if it is adequate for only 90% of them it's a huge threat to the big guys. There will always be a niche for the big guns that handle huge databases with many, many simultaneous updates. But that's a small fraction of the total universe of database installations and I don't think it's enough for the big guys to continue to be profitable.

Re:Version 4 Will Tell

[RevAaron]

Considering the way MySQL is (ab)used, flat text files, serialized data structures/objects or XML files would be very adequate and just as convenient for what 30-40% of what MySQL is used for. Mind you, they would be sufficient for what 30-40% of the big boys of enterprise databases are usually used for, but they're often applied when they're actually needed.

That isn't to say that I am against using databases, but the overhead of MySQL is often pretty absurd for very simple dynamic websites (hell, a lot of kinds of dynamic web sites) and desktop apps managing a relatively small amount of information. If a DB was integrated into the OS as the preferred method of storing data, with the overhead paid for across many apps in increased convenience, it'd be worth it. But why the hell should I need to install MySQL just to maintain a list of todos and contacts? Look on Freshmeat- there is a torrent of applications using MySQL for managing small amoutns of data, both web and desktop apps.

It's too bad most Linux developers aren't interested in doing something really forward-thinking. If there was a DB integrated into the OS, and apps encouraged to use it, with avenues of data management made easily available to the user, computing could be actually pushed ahead by Linux. But not today, and probably nor ever.

Oh yes, my point: most of these apps would do fine with a flat file or (if one must get fancy) an XML file to manage this data.

Re:Version 4 Will Tell

[walt-sjc]

Let me expand on that :-)

The bottom line is that Oracle (and other enterprise DBs like MSSQL, DB2) have enterprise features that MySQL (and other open source DB's) just don't have, and probably won't have for Years. It's more than just the ability to scale, and raw performance on simplistic tables.

Most everyone is very much aware of MySQL and other open source solutions are great for certain types of applications, but horrible for others.
That said, many enterprise users use Oracle in cases where MySQL would be much more cost effective, and probably better performing as well.

Re:Version 4 Will Tell

[walt-sjc]

That's a good question for IBM who did exactly that back in the early 80's with the AS/400. In fact, the entire architecture was designed around it.

Re:Version 4 Will Tell

[Compuser]

BeFs did this and AFAIK Reiser is working on this
for his FS. And besides, MS has promised this for
their future Windows, so you know it will be
copied and made mainstream eventually.

Re:Version 4 Will Tell

[RevAaron]

BeFS didn't do it. BeFS had attributes, which was a definate step in the right direction.

The Newton OS did it, with an object database. Down the line, other PDA OSes did it as well- Palm OS and the Helio's VT-OS both provided a database as the only means of data persistence. The Palm OS and VT-OS DB systems are quite a bit more restricted than the Newton OS's OODB or the theoretical system-wide relational DB we're discussion.

Dynapad (my PDA OS/OE) take an approach similar to the Newton OS with an system-wide object database.

Re:Version 4 Will Tell

[the eric conspiracy]

MySQL is very adequate for 99% of all users including most large enterprises and certainly most websites.

I used to be a big fan of MySQL, mostly because it was moderately capable and free. Now that I have tried Postgres though there is no way I would go back.

Yes, version 4 will be an improvement, BUT it is still missing many key features like views, triggers, full outer joins, update with subselect, that are already present in Postgres, and the fact is I've been using the features that MySQL is promising for the future for a year and a half now.

The following site does a very good comparison between the feature sets of MySQL, Oracle and Postgres.

http://det-dbalice.if.pw.edu.pl/det-dbalice/docu me nts/all/html/db_compare/db_compar_chp01.html

Re:Version 4 Will Tell

[RevAaron]

I don't specifically mean the kernel. That would seem a bit unnecesary, unless it provided a pretty big boost in performance and was universally used by applications and the system alike to make it worthwhile.

The point would be for a unified model of data format, access and storage. No more file format worries. Empowering users to manage and manipulate their data. Easy sharing between apps on the same machine, over the network, across platforms.

The important change isn't in capability but in the way of doing things. Since I do not mean stuck in the kernel when I say OS integration, I simply mean that it would be a core part of the OS used by all applications. Instead of files as we know them. This could be provided by an existing user-space solution, but until there is some standardization the benefits wouldn't really materialize. E.g., it doesn't matter if MySQL is installed on this Linux box on my desk if none of the applications use it.

Re:Version 4 Will Tell

[captaineo]

I agree that 90% of people who use databases would really be better off with simple text files or an in-memory server... SQL is overkill for shopping lists and username lists, though it certainly makes sense for high-end, high-performance stuff.

ReiserFS is trying to become a database. The first step was making ReiserFS v3 perform well on small files. (the idea is to store tree-like/XML-ish data in a directory structure where each file contains only a few bytes of data). Last I heard, ReiserFS v4 was moving towards a full transaction API that will allow atomic batch transactions on more than one file in a single system call. (aside from the space overhead problems inherent with storing tiny amounts of data in each file, most current filesystems don't allow atomic transactions that involve more than one write to one file...)

Re:Version 4 Will Tell

[-eddy]

If a DB was integrated into the OS as the preferred method of storing data...

Isn't that what Microsoft calls the Registry?

Re:Version 4 Will Tell

[rycamor]

Yes, version 4 will be an improvement, BUT it is still missing many key features like views, triggers, full outer joins, update with subselect, that are already present in Postgres, and the fact is I've been using the features that MySQL is promising for the future for a year and a half now.

Not to mention column and table constraints, stored procedures, extensible datatypes, user-defined operators, query rewrite rules, and schema and domain support.

Re:Version 4 Will Tell

[foniksonik]

I was under the impression that 4 had already been released... is this not so?

What gives? Personally I only use 3.x.x but that is because the scripts I use were developed for it and don't take advantage of 4 AND I'm not as comfortable with 4, what it brings to the table and what it may or may not break in my setup.

Re:Version 4 Will Tell

[JPriest]

You mean like Longhorn's SQL based WinFS?

Re:Version 4 Will Tell

[Anonymous Coward]

Peformance isn't what matters. No one cares about the MySQL overhead. SQL is a very *convenient* abstraction and way to access your data. It is easy to setup, and then to process queries. Even if people abuse it, it works. And for 90% of these applications working is good enough.

Re:Version 4 Will Tell

[HydroCarbon10]

Not to mention that the current version seems to have a mysterious problem with randomly melting databases.

I'm bookmarking this story so next time my database melts I can come here for a quick laugh before going off to repair it.

Re:Version 4 Will Tell

[lseltzer]

no, the registry is not a generalized database, it's a proprietary hierarchical store. Rhetorically you could call it a database, but you could call anything a database.

Re:Version 4 Will Tell

[SamBeckett]

I put on my databases this morning. The dog kept on biting my databases. After I got home this evening, I took off my databases and they were stinky. Luckily, I had some spare databases around to wash them with.

Re:Version 4 Will Tell

[Anonymous Coward]

larry ellison has been talking about developing a
DB with an integrated OS for some years, to minimize overhead on DB servers. Sort of the flip side of what you're talking about.

ls

Re:Version 4 Will Tell

[KilerCris]

This is the point I was about to make. I've worked with flat file , serialized data, xml files, and MySQL for databases for many small websites. Sure the first 3 are faster in execution, even if it isn't noticable, but using MySQL shaves a lot of time off the coding.

Re:Version 4 Will Tell

[RevAaron]

Yup. I can't say I keep up on the state-of-the-art in future MS OSes, but the wee bit I've read on OSNews sounds like it may be something good. Let's hope it works out and is useful...

Re:Version 4 Will Tell

[RevAaron]

As someone else pointed out, the registry is a kind of database yes, but it is not what I'm talking about. The registry isn't used for- and wasn't designed for- all-purpose data storage. The registry is used for storing preferences, and is limited to storing strings, bytes and keys (IIRC). No BLOBS, among many other missing features. You query it just by specifying a path, no query language or query language surrogate.

Re:Version 4 Will Tell

[Anonymous Coward]

There are two camps of thinking on this: the flat file people, and the relational database people. I think I'm from somewhere in between.

Much of the time, someone starts working with flat files and ends up wasting a lot of time writing, testing and debugging code that simply does the file handling work. Sure, a database may be considered "overkill" for some tasks, but most of the time it's worth it just for the time you save not worrying about that extra layer.

Plus, it is a heck of a lot easier later on to:

- modify the database schema. eg. Even if your specific program doesn't need additional columns, you can add some which will be used by another program. XML can help alleviate some of that, but it's bloated so using a database instead really isn't really any more overhead.

- multiple user access

- security!

- easy to backup/restore

- ACID properties

- multi-language support

- so the next person to be in your position will know what the hell is going on. :)

- tons of other things you'll say "oh yeah" about down the road when your flat file system code doesn't cut it anymore.

Where are flat files good? Client application settings, documents to exchange with users, simple and single-user data.

Re:Version 4 Will Tell

[RevAaron]

but using MySQL shaves a lot of time off the coding.

Only if you're doing everything manually. The times where I've gone with serialized data or an XML file, the code ended up less than the same solution done with an SQL connection. However, if you're doing this with a language or library that makes you reinvent this wheel every time, yes, you're worse off in terms of time spent coding. But even if you're stuck on a platform like such, there is no reason you cannot make a relatively small wrapper around these operations which redunancy and amount of code needed to do these things. All the benefits, none of the negative aspects.

Re:Version 4 Will Tell

[RevAaron]

I'm inbetween those camps as well- if you need real relational database features, use one. A lot of what I see MySQL being used for does not need this. Will not suffer from a minor schema change down the line, etc.

Re:Version 4 Will Tell

[Chester K]

It's too bad most Linux developers aren't interested in doing something really forward-thinking. If there was a DB integrated into the OS, and apps encouraged to use it, with avenues of data management made easily available to the user, computing could be actually pushed ahead by Linux. But not today, and probably nor ever.

An application integrated into the OS? You mean like Internet Explorer?

Re:Version 4 Will Tell

[Tony-A]

but the overhead of MySQL is often pretty absurd for very simple dynamic websites (hell, a lot of kinds of dynamic web sites) and desktop apps managing a relatively small amount of information.
You're right, but "high overhead" is not what I ever expected to see in conjunction with MySQL. Your point is well made. I would hate to see init (or anything more "basic" than MySQL) having to depend on it.
That point also explains a lot of the success of MySQL. For a lot of things, MySQL is more than enough. The key thing to remember is that slow readers and fast writers is not what MySQL is designed for. If you can live within MySQL's limitations, I don't think anything else will touch it, even at the enterprise level.

Re:Version 4 Will Tell

[lcracker]

I don't have a lot of experience with using it, but you're wrong.

It supports a few more types than that, and it *does* support BLOBs to store whatever else you want.

You can enumerate "folders" (keys which reference a structure that hold other keys) and "files" (keys that reference values) in the registry separately.

Re:Version 4 Will Tell

[timeOday]

If a DB was integrated into the OS as the preferred method of storing data, with the overhead paid for across many apps in increased convenience, it'd be worth it. But why the hell should I need to install MySQL just to maintain a list of todos and contacts? Look on Freshmeat- there is a torrent of applications using MySQL for managing small amoutns of data, both web and desktop apps.

I don't understand; it seems that you are getting your wish of pervasive database usage, but are unhappy about it. Doesn't the torrent of apps mean it's becoming the preferred method?

Re:Version 4 Will Tell

[SenorMooCow]

Version 4 is currently in "gamma" release, in other words very stable beta. I use it on my website with no problems (although it does not have the load that a large website would need to stand up to).

Re:Version 4 Will Tell

[Arethan]

Actually, this is an interesting point, but it isn't as valid as you make it out to be. There is absolutely no good reason to embed an RDBMS into an operating system. Not at the level that you are referring to anyways. If you mean filesystem attributes, then that is a different fish, but making a database part of an OS is a bit much.

XML and flat files are good for data that does't change very often, and are only ever edited by a single user. (XML is also a good way to feed data fram a database into an application) Beyond that, they are pretty useless. They require far too much time on the developer's part when data within given contraints are needed, and coordinating updates between multiple processes can easily turn into a nightmare. Not to mention scalability. XML can get pretty large, and due to the non-indexed nature of the data, it can take a long time to read through it all looking for what you need. For many applications, using a database just makes sense.

The whole point of an app using a database is to offload the storage specifics onto another program. I applaude the developers of these "MySQL (ab)using programs" for making the decision to focus on their products features and stability, rather than on how they will store their records. Besides, once you install MySQL (or Postgres) once, then it is there for all of your DB dependant apps to use.

Not to mention the fact that MySQL 4 has standalone features that make your argument pretty moot. Any application can link in the MySQL core at build time, and will be able to have it's own MySQL databases separate from any active system wide MySQL instance. This gives developers an SQL storage system, without requiring the user to install a database on their own. It just comes bundled with the app, and when the app loads, the MySQL core is loaded right along with it.

I'll keep using databases for storage, and flat files for configuration data, thanks though! ;)

Re:Version 4 Will Tell

[einhverfr]

I used to be a big fan of MySQL, mostly because it was moderately capable and free. Now that I have tried Postgres though there is no way I would go back.


That makes at least two of us.


Yes, version 4 will be an improvement, BUT it is still missing many key features like views, triggers, full outer joins, update with subselect, that are already present in Postgres, and the fact is I've been using the features that MySQL is promising for the future for a year and a half now.

My big complaint with MySQL is that it originally let me use very sloppy database design which limited portability of an original application. However, I still so support it, and have this to say aside from that.

I do highly suggest always running MySQL in ansi mode because that maximizes portability of your applications. Referential Integrity Enforcement is probably not needed for a majority of small databases, and can be ignored. But for business-critical databases it is absolutely necessary. So until constraints and triggers are included, I will not consider it to be sui9table for more major applications where the data is important.

For example, MySQL is OK for many many web sites, but I would probably avoid using it for ecommerce at this time. And although my CRM suite supports MySQL, this was because I had as many concerns with PostgreSQL as a prototyping RDBMS before 7.3 came out. Of course, by the time MySQL 7.3 is available, it will be quite a database manager! ;-)

Re:Version 4 Will Tell

[RevAaron]

Hrmm, perhaps- I made that statement based on what RegEdit lets you do and what one can do on WinCE. (I've done some WinCE programming, but nothing on desktop Windows)

All the same, the Registry is still not what I mean- after all, if the registry really puts itself entirely in RAM, causes massive instabilities, and is generaly flaky it couldn't be used as an all-purpose general storage mechanism for the entire OS.

Re:Version 4 Will Tell

[RevAaron]

Yeah, those two different points were pretty weird to make in the same post. :P

But no, I don't see a lot of random apps all happening to use MySQL as being pushing toward this goal. Without a level of standardization- on the database, availability to users, methods for the user to easily manipulate and access the data- it's irrelevant. The handful of apps which use MySQL for storing simple data isn't pervasive. When you have just a couple apps on your system doing this, it doesn't make sense. If all or most of them were, the benefits would be a lot more real.

It's not that I'm unhappy about these apps, I just think it's a big goofy unless a lot of apps adapted to this. I guess it's a bit silly of me to want this- a very slow and gradual adoption is much more likely that some magic conversion. At least on relatively stagnant OSes like Linux. (No offense/flame meant, simply that it embodies old ideas. These old ideas work well enough, but they're not getting any better.)

Re:Version 4 Will Tell

[oconnorcjo]

Considering the way MySQL is (ab)used, flat text files, serialized data structures/objects or XML files would be very adequate and just as convenient for what 30-40% of what MySQL is used for.

...

It's too bad most Linux developers aren't interested in doing something really forward-thinking. If there was a DB integrated into the OS, and apps encouraged to use it, with avenues of data management made easily available to the user, computing could be actually pushed ahead by Linux. But not today, and probably nor ever.

I find it strange that you say MySQL is more overhead than needed for most projects but then you turn around and say it would be cool to build a DB into Linux.

DB's are wonderfull because they are designed to make it easy to organize and modify data. I am apt to use a DB even for "simple things" because flat files are slow and are not as easily adaptable to future needs (and the convenience of a DB far outweigh the overhead). However, the idea of DB being built into the Linux Kernel, makes me shudder. Each DB has its place and none of them belong in kernel space. The Linux Kernel developers follow KISS and plugging a DB into an OS is not KISS.

What you might be better off with is if Gnome and KDE made Postgress or MySQL (or some other XYZ DB) a component of thier respective Desktop Enviroments and wrote Gnome/KDE interfaces and API's to facilitate integration of other programs to use the DB. That is a better solution to what you are thinking.

Re:Version 4 Will Tell

[IamTheRealMike]

The Palm OS and VT-OS DB systems are quite a bit more restricted than the Newton OS's OODB or the theoretical system-wide relational DB we're discussion.

If you read Hans Reisers whitepaper on the future of ReiserFS, you'll find he puts forward some convincing arguments (at least to a layperson like me) as to why using a relational database at the filing system layer would be a bad idea. The central theme is that of eliminating information by imposing too rigid a structure on it (or something like that). I think he intends for the final result to be based on set theoretic semantics mostly (although with a variety of different layers, so you can join together a heirarchical namespace to a set theoretic namespace).

So, people are indeed working on doing what you wanted, and have been for several years, but it's tremendously complicated and performance concerns have required a lot of groundwork to be laid first.

Re:Version 4 Will Tell

[darnok]

> That said, many enterprise users use Oracle in
> cases where MySQL would be much more cost
> effective, and probably better performing as well.

It's remarkable how many shops will choose Oracle, DB2 or SQL Server over MySQL or Postgres for even the most mundane tasks, and choose to pay HUGE licencing fees as a result.

One project that I worked on had the (Web-based) clients sending out parallel requests to a mainframe DB2 backend, and progressively assembling the results in an "intermediary" SQL Server database. Once all the results from the DB2 database had been collated together, the data would then be sent from the SQL Server database back to the client. Although this wasn't a particularly high-throughput app, the sheer quantity of iron that was required for the SQL Server DB boggled the mind - I think it was something like a 4-way active-active cluster - and the SQL Svr licence fees were something over $100k - **for a database that only ever held transient data**!!!

MySQL would have been perfect for this job. It would have saved big bucks on licences alone, then it would've saved more due to its much faster throughput and lower resource requirements. If really necessary, it could've run on hardware bigger than the biggest Intel boxes, which obviously wasn't an option with SQL Server.

Why didn't they go with this? Three words - Microsoft Consulting Services.

Re:Version 4 Will Tell

[indiancowboy]

Well, If 90% is the market for non-critical, non-transactions intensive sites then is MYSQL really the one to take on this 90% ? Hello, is'nt someone forgetting MS Access ? The small guyz(sites) anwyas would be able to make do with windoze. So is'nt access the more obvious choice? Also considering windows having a huge market share in the server segment(as of today atleast!), I think access would it.

Re:Version 4 Will Tell

[grahamm]

No not an application but filesystem interfaces. As an example look at the ICL mainframe VME/B Operating System. This puts the record access mechanisms in the OS and applications access the files at record level rather than byte level.

Re:Version 4 Will Tell

[skillet-thief]

Yes, version 4 will be an improvement, BUT it is still missing many key features like views, triggers, full outer joins, update with subselect, that are already present in Postgres, and the fact is I've been using the features that MySQL is promising for the future for a year and a half now.

One thing that probably keeps a lot of users (esp. web people) loyal to MySQL, is the fact that they learned SQL on MySQL and don't really know what else it should be doing for them.

In fact, that is where I am right now: just realizing the limits of what MySQL can do and tired of writing various hacks via PHP or Perl to get around certain weaknesses.

But I think that a lack of SQL culture could keep many users locked into MySQL when other DBs might be better for them.

Re:Version 4 Will Tell

[gilesjuk]

Stored procedures make things easier too, many wont touch it until then.

But isn't MySQL best suited to being a lightweight database for webservers anyway?

Re:Version 4 Will Tell

[let_freedom_ring]

"It's too bad most Linux developers aren't interested in doing something really forward-thinking. If there was a DB integrated into the OS, and apps encouraged to use it, with avenues of data management made easily available to the user, computing could be actually pushed ahead by Linux. But not today, and probably nor ever."

Actually, ReiserFS is pursuing this goal by pushing the DB storage layer into their file system implementation. The file system is a natural place to put this as it (unlike the rest of the kernel) provides persistent storage.

Re:Version 4 Will Tell

[pclminion]

Not to mention column and table constraints, stored procedures, extensible datatypes, user-defined operators, query rewrite rules, and schema and domain support.

How about views? All I want is friggin' VIEWS. Why is that still missing?

Re:Version 4 Will Tell

[tellezj]

Just a comment on your sig. There is no comma between "nation" and "under". It should read "one nation under God". The difference is subtle but important.

Re:Version 4 Will Tell

[4of12]

be better off with simple text files or an in-memory server.

Absolutely.

With the trends in memory pricing and the anecdote of Google's excellent performance being due to keeping so much in memory, I think we'll see the trend to using in-memory db's growing rapidly, whether it be something new, or some increased use of buffer caches in any of the current SQL offerings: Oracle all the way down to MySQL.

Re:Version 4 Will Tell

[jeremyp]

XML and SQL are designed for different jobs. XML is a language for describing structured data, nothing more, nothing less. SQL is a language for manipulating structured data in a flexible way. I see no reason why you couldn't issue an SQL query to a database and have it return the results as XML. Well, there is one problem: PHP (for instance) has a nice little function call that will take a row from the mysql result set and return it as an associative array indexed by the column names in native format. Other languages have equally simple ways to get the data in a usable form. Why mess around with XML parsers?

SQL comes into its own as soon as you have more than one logically distinct but related set of data. SQL allows you to query that data in arbitrary ways not necessarily catered for in the original design.

Further, XML is a poor way to store data unless you intend to read it all into RAM before doing anything to it. An XML file is essentially a stream of character data which means that it is difficult to index it (making searches slow), difficult to insert records into it (well you could just append new data thus making searches even slower) and not space efficient (lots of extraneous syntax which is only there to make it easier to export to other apps + binary data has to be stored encoded as character data in some way which invariably makes it bigger).

Re:Version 4 Will Tell

[NitroWolf]

I wish I could switch to Postgres ... the problem is, though, that a lot of the applications that fall into the 99% range as mentioned above only work with MySQL, or possibly MySQL and MSSQL... not Postgres :( At least the applications I use, don't...

Which is sad. I would switch to Postgres for the subselects if nothing else.

Re:Version 4 Will Tell

[Nyarly]

Honestly, the work I've seen XML do well is as follows:

1. Document storage. Like HTML, only moreso. XML is great for storing documents that are intended to be human readable - both in their raw format and in their eventual display. Word processing, presentations, web pages, that lot.
2. Configuration and commands. It's nice to be able to have a single form of configuration, without having to guess which bizarre flavor of syntax is going on. Do spaces count? With most /etc files, who can say? (Hopefully man -S8, but sometimes you aren't so lucky.) With XML, you know the syntax rules, and if you can find the DTD (or XMLSchema), you at least know the words and the grammar, if if you have to guess what you're saying. (On the other hand, OpenNMS' "IPLIKE" tag gets several demerits for lazy coding and XML abuse.)
3. Serialized communication and object marshalling. My only solid example on this front is Jabber, but the theory is quite sound. There are plenty of serialize-to-XML packages running around, as well. Granted, a little application level compression would probably help. Hell, gzipping the packets before sending would probably make up in network time what it eats in overhead.

Further, XML is a poor way to store data unless you intend to read it all into RAM before doing anything to it.

Cue pedantry theme. Strictly speaking, SAX doesn't require XML to be loaded into RAM. But to do anything useful, it does require that you load it through RAM, so you're point about lacking of index is well made. SAX is great for the applications listed above. Document editing is better served by some variant of DOM, in which case, the difficulty of inserting records dissapears. Size issues could be easily corrected, if it were an issue, with simpel compresion, since XML naturally includes many repeating strings.

However, your point that XML and RDBs are seperate entities is well taken, and with that, I'll strongly agree. They have very different roles. Which doesn't stop the one being used in the other's place with alarming frequency.

Re:Version 4 Will Tell

[Anonymous Coward]

I was wondering when someone would chime in the 400. IBM yes did do this and front-ended it with an interface called DB2/UDB. There are SO many DB-like features in the OS like primarily the process TABLE.

Re:Version 4 Will Tell

[WNight]

ReiserFS v4 is looking to have a transaction interface. This initially will just be accessible by the kernel to allow out-of-order write caching to best utilize the disk.

(User writes block 1 and 2, block 2 is nearby, but block 1 isn't. It's faster to write 2, then 1, but if you crash between them your data is in an inconsistent state... With transactions you can safely write 2 first, then 1, because you know that they'll both happen, or not, as a group.)

But, to be a database replacement you need to allow independent transactions without blocking. For instance, someone runs a really complex select on a huge DB at midnight. It runs for three hours before returning an answer, but it sees the DB as it was at midnight, without any of the data written since then. But, you can't block writes while this happens...

If ReiserFS does manage to allow this, it'll be incredible. I doubt the default filesystem will support it. It's the kind of thing that you rarely need in filesystem, but costs a lot to provide.

Re:Version 4 Will Tell

[Corporate Gadfly]

The following site does a very good comparison between the feature sets of MySQL, Oracle and Postgres.

http://det-dbalice.if.pw.edu.pl/det-dbalice/docu me nts/all/html/db_compare/db_compar_chp01.html

An otherwise informative and to the point post has been rendered useless by you not using the A HREF tag. Copying-and-pasting URLs wastes time (not to mention the embedded spaces to weed out)

Here's the same Comparison of Oracle, MySQL and Postgress DBMS as mentioned in the quoted comment above.

Re:Version 4 Will Tell

[WNight]

I didn't get the idea that the parent posted wanted to shove MySQL or Postgres code into the kernel. I thought they wanted to make a common interface available so that having an RDBMS available was as common and expected as having a /tmp directory.

Today, you can assume perl is on a system and there's 95% chance of python being available. But for an app developer that's not where it ends. If you want a database available you have to get your install scripts to install Postgres, or get the user to do it, and then cope with all these different directory structures, a real nightmare of configuration options.

What if there was a simple library call 'sql-query-to-systemDB' and the distro maintainers made sure that there was an rdmbs available and they took care of the back-end, like is it mysql or postgres, etc. Your application would just create tables and query away without any more worry than creating a temp file or allocating memory.

Now we'd scoff at an OS that didn't have tcp/ip networking available. Virtual memory is a requirement. When will we start seeing RDMBS availability this way?

Re:Version 4 Will Tell

[WNight]

This is why Oracle ported to Linux. They get the benefits of having their own OS (customization) and their customers don't have a pay a second set of license fees.

Many developers of destination-applications (something a customer demands, and buys a system to run) like Oracle, 3d rendering, etc, are looking at Linux. It lets them install onto an empty system without worrying about changing OSes, security holes opened by other processes, etc. It also lets them provide things like the ability to remote configure the computers with proven technology like ssh and X instead of writing their own tools.

Next to a copy of many applications, the hardware is a trivial cost. $15k for a 3d package, $2.5k for a smoking x86 computer to run it on.

And, these companies analyze the market and figure that a customer will be willing to pay $40k for a system with Oracle handling n transactions per second. If they can remove the OS cost, more of that $40k goes to them.

Re:Version 4 Will Tell

[Ryosen]

Perhaps he was referring to MS Access, the binaries for which are included in the Windows distro.

Re:Version 4 Will Tell

[Omega996]

doesn't UDB refer to the versions of DB2 that run on Open Systems (aix, solaris, hp-ux, linux) and windows nt derivatives?

Re:Version 4 Will Tell

[the eric conspiracy]

and postgres i would have to be careful especially about performance.

I have found that postgres requires a lot more attention to tuning than does MySQL, especially in terms of memory buffer space. However I have found that there are cases where postgress outperforms MySQL - particularly in the area of inserts because of the differences in locking features.

I think PostgreSQL is more of a threat

[srn_test]

We used to use mySQL, but moved to postgreSQL for performance reasons, and we're glad we have.

On the postgreSQL general mailing list, people rarely talk about mySQL anymore (let along mSQL). It's (mySQL) is generally regarded as a good alternative to the Berkeley DB stuff (i.e. non-relational), whereas postgreSQL these days gets lots of traffic from Oracle people wanting to go somewhere cheaper.

Oracle mustn't be happy, I'd think.

Re:I think PostgreSQL is more of a threat

[j3110]

I agree.

I've developed my last application for MySQL. Everytime the server looses power, I have to ssh into client's servers and tell them how much data they've lost (repair table). It's not a happy time, and buy a UPS is not reassuring (most have them, accidentally bumping power switches/knocking cables loose still happens).

InnoDB doesn't have this problem, but then again, it has buggy key problems on all of my servers. Sometimes it can't find a record that is there (often this is worse than just loosing the record... you can't create a duplicate). I have to periodically rebuild the index on inno, so I scrapped it too.

I've used postgreSQL before, and it seems MUCH more robust. It can be a little slower on certain queries, but I'll sleep better after my clients are ported over. I also get updatable views, custom objects, sub-selects, embedded procedures (in a variety of languages), transactions, cross table deletes/updates, and speed when I take advantage of key clustering and these other features rather than hacked solutions for sub-select. I've seen people select "delete from table a where bid=\""+b.id+"\"" from b where c=3 into outfile d; then run mysql -u user -p d

Sure vacuum could be automatic when there is a great degree of fragmentation, but those are scratches compared to the gaping holes in MySQL like ACID compliance. I've looked at version 4, and it changes the queries so much that I would have to port my app to version 4. It's best I go another route than be disappointed again.

It sure does help Oracle migrants that pgplsql is about the same as Oracle's plsql :)

Re:I think PostgreSQL is more of a threat

[vandan]

I had to do some work on a Postgres system over the weekend. I come from a MySQL background.

Having heard people (a handful of noisy Postgres users I suspect) rant about how much better Postgres is and how unfair it is that MySQL is the No1 open source database, I had high hopes for my first experience.

I was quite disappointed. There are bugs. And those bugs have bugs. And the bugs' bugs have bugs too. You can't import from a backup made by a different verion of Postgres. That was my first problem. The older version I was using was trying to parse the comment marks (---). OK. So I upgraded to the latest version. The GUI (pgadmin) didn't work. I frigged with it for a while and decided to give up and get back to work. The command-line psql segfaults. Now that is a SERIOUS problem. I have never seen the MySQL client segfault, and I have used it a lot. Then there are stupid, stupid things like not being able to delete a column, or change a column type. Having used MySQL, I took those things for granted. I am always changing a column type to varchar(x) and pruning garbage off it (eg dollar signs and commas) and then converting it back to a numeric (double) field. Can't do it in Postgres. Why? I assume it's because all the developers have since migrated to MySQL.

Now I understand that Postgres supports views and stored procedures - both of which I'm not particularly impressed with. There are very easy work-arounds that you can do on the client side (which I do here at work), such as put all 'views' in functions, and just call the function to initialise the SQL string you want. Not very hard.
As for the relational stuff, InnoDB has supported that for ages and ages. Don't give us that non-relational argument crap no more.

Why do people use Postgres? I don't know. Maybe there are some who really really really need triggers. God knows why. As the MySQL developers point out (and they are quite correct) - this sort of thing should NOT happen on the server; it should be contained in the client software or middle tier. Views? See above. Don't need them. Stored procedures? Handy but honestly can be done just as easily at the client end.

Having used both, I can honestly say that I'd rather wait for MySQL to add the one thing I'm actually waiting for: stored procedures, than use Postgres. Postgres just gives me the shits. Maybe that's why MySQL is the No1 open source database.

Re:I think PostgreSQL is more of a threat

[maraist]

Actually in 7.3 you can finally drop columns, fwiw.

I don't know about MySQL 4, but the biggest problem is that MySQL seemed to emphasise speed over robustness.. Large scale benchmarking in the 3-tree caused corrupted databases. postgres has had a journal for a little bit now (don't know how long though). Plus postgres's version-based rows allows for some really high performance parallel transaction control.

It all comes down to what you are really trying to do.. If you're willing to lose a day's worth of work, and you're not going to have more than a couple dozen simultaneous connections, then MySQL is probably good for you.

Perhaps 4 has allowed MySQL to catch up with Postgres / surpase it, but it's had too murky a history for a lot of businesses who rank data-integrity number 1.

Re:I think PostgreSQL is more of a threat

[leonbrooks]

You can't import from a backup made by a different verion of Postgres.

I can. I have, many times.

The older version I was using was trying to parse the comment marks (---).

The command-line psql segfaults.

These lead me to suspect that your implementation was broken. I've never seen them happen.

not being able to delete a column, or change a column type.

ALTER TABLE [ ONLY ] table [ * ] ALTER [ COLUMN ] column { SET DEFAULT value | DROP DEFAULT }

ALTER TABLE [ ONLY ] table [ * ] RENAME [ COLUMN ] column TO newcolumn

I am always changing a column type to varchar(x) and pruning garbage off it (eg dollar signs and commas) and then converting it back to a numeric (double) field.

Hey, what? You can't store dollar signs and garbage in an integer or float, you shouldn't be trying to feed that to your db in the first place! If you want to do that kind of thing, amidst a `live' table is not the place for it: use a temp table and do it properly. Whis is probably why the PostgreSQL people didn't implement it.

I'd rather wait for MySQL to add the one thing I'm actually waiting for: stored procedures

I hope you've got a lifespan like Methuselah's, then. PostgreSQL does stored procedures in a variety of languages already. Your post does sound like a BASIC programmer grunting and squealing when presented with a real language that insists on him doing stuff like decalring variables, and has scoping etc, forcing him to do `work' (actually investing in manageability) that he didn't have to do before - at least, not up front and in small doses.

Re:I think PostgreSQL is more of a threat

[kootch]

i think you really need to convince your client to move their server out of the hallway closet and into a hosting facility...

the server loses power? people knocking out the power cords? i really hope this isn't a mission critical system ;)

Re:I think PostgreSQL is more of a threat

[synx]

it happens, i work for a major company, and due to a wiring error and a mistake made by power techs, major switches and about 20 servers went down, taking out some major production systems.

So no matter what happens, your database will eventually fail and lose power. Even if the power is 100%, the db software and/or OS will crash.

And yes, oracle crashes :-) And hard. But it comes back up cleanly, which is the important part.

Re:I think PostgreSQL is more of a threat

[Ragica]

Background vacuuming is on the TODO list. You're probably aware that as of version 7.2 vacuuming no longer requires table locking, so it seems quite feasable to vacuum in the background to me. There is also a 3rd party daemon available which will automatically vacuum tables based on a configurable activity threshold (i've never tried this though).

Re:I think PostgreSQL is more of a threat

[jasondlee]

At my company, we even had a UPS tech forget to configure the UPS correctly when he was changing out some fans on the machine. He flipped the switch, and there went our machine room. He took out a couple of SQL Servers, and handful of Linux boxes, some random Windows servers, and a huge iSeries 820 (AS/400). Nothing is 100%.

jason

Re:I think PostgreSQL is more of a threat

[Heikki_Tuuri]

Hi!

Can you describe the index problem in more detail? Please send a bug report to mysql@lists.mysql.com.

- What MySQL version did you use?

- Did CHECK TABLE report the table ok?

- What kind of SELECT queries did you execute and what did they report?

- Are you aware that in the AUTOCOMMIT=0 mode you have to COMMIT your read transaction to advance the consistent read timepoint? Some users do not know this and then wonder why committed data is not visible in another connection. InnoDB serializes read-only transactions at a precise timepoint so that all consistent reads are really consistent with respect to each other. The default transaction isolation level is therefore called REPEATABLE READ.

Best regards,

Heikki Tuuri
Innobase Oy

Re:I think PostgreSQL is more of a threat

[Lumpy]

Everytime the server looses power, I have to ssh into client's servers and tell them how much data they've lost (repair table). It's not a happy time, and buy a UPS is not reassuring (most have them, accidentally bumping power switches/knocking cables loose still happens).


this is not the fault of the DB this is the fault of the data-owners.

Obviousally your clients have very little value to their data as they do not have the proper equipment to protect it. Power switches getting bumped? what are they running on someone's old dell desktop in a closet? Buy a Compaq ML530 or other real server instead of running on junk. I have to hold in the power switch for 10 seconds before it starts a shutdown because of the configuration I have. a full power crash means unlocking the back of the server rack and unplugging the 3 power cords from the power supplies. As for UPS, I have a APC1400 rack mount for each server and the batteries are replaced YEARLY as well as the UPS's tested weekly.

we value our data, and it costs us $10,000.00 for every day of lost data. so spending money to protect that data is important to us.

Obviousally your clients do not have a high cost or high value to their data or they really dont care much about data loss, otherwise they would invest in server class hardware and other assurances. I can only imagine the nightmare that is their backup solution if they have such low quality IT equipment in place that your stated problems actually affect them.

Re:I think PostgreSQL is more of a threat

[ChannelX]

This response is pure trash. I've never actually used PostgreSQL but I can tell you why this response is pure trash:

"Maybe there are some who really really really need triggers. God knows why. As the MySQL developers point out (and they are quite correct) - this sort of thing should NOT happen on the server; it should be contained in the client software or middle tier. Views? See above. Don't need them. Stored procedures? Handy but honestly can be done just as easily at the client end."

Written by someone who either doesn't care about proper programming techniques or has no clue about them. If the MySQL developers are putting out some trash that say things like triggers should not be in the database but in the client they deserve no respect. They have no clue what they're talking about.

The poster needs to learn the reasons behind using stored procedures. Sure the same thing can be done on the client. The client doesn't offer any of the advantages of stored procedures however: separation of model/view/controller, precompiled for speed, etc.

What a ridiculous posting.

Re:I think PostgreSQL is more of a threat

[jeremyp]

this is not the fault of the DB this is the fault of the data-owners.

A commercial DB server (i.e. one used commercially not one that costs money) should be able to survive something like a power failure without corrupting data. There are many reasons why a server might go down even when apparently fullly protected by UPSs etc. Some I have seen are:

- UPS circuit breaker being tripped by idiot maintenance man who was cleaning bits of the UPS with a metal brush

- kernel panics

- accidental shutdown by tired, stressed and overworked sysadmin who was logged into the wrong server when he issued the "halt" command

- SCSI devices locking up the bus

- database application failures (does not bring the machine down but can cause referential integrity failure without a transaction based DB).

All that hardware you are using is great (essential) for 24x7 availability but overkill if you accept lower availability but still need consistent data.

Re:I think PostgreSQL is more of a threat

[Lumpy]

Ok I'll answer these one by one....

- UPS circuit breaker being tripped by idiot maintenance man who was cleaning bits of the UPS with a metal brush



Why?? why does ANYONE other than the IT professional responsible for it have keys to the server rack doors, both FRONT and BACK. soory, this is negligence on the owners and maintaince personell.

- kernel panics

I havent had a kernel panic on a critical server for over 4 years now. what are you doing running bleeding edge kernels on a critical server? otherwise what are you doing running services other that what is needed on a critical server that might be able to crash the kernel? other than NT the Os is rock solid, and in NT's case Nt4.0SP6 on good hardware is rock solid.

- accidental shutdown by tired, stressed and overworked sysadmin who was logged into the wrong server when he issued the "halt" command

Why? was it a kernel upgrade? there is NO reason for a server halt other than hardware or kernel upgrade (Again only the case if you run NON Microsoft OS, those you have to reboot if you fart near it) and even then it's a dumb way of shutting down a server. SHUTDOWN is the proper command and anyone that uses HALT on a regular basis needs a good talking to and possibly revokation of root access. Finally, making excuses for blatent mistakes are worthless. that overworked sysadmin needs to pay attention. I along with 3 other admins here have pulled 36 hour shifts during a crisis or special project. if you don't doublecheck yourself when doing IMPORTANT things that can cause serious trouble then it's either ignorance, apathy or laziness taking effect. Again, the fault of the operator/owners.

- database application failures (does not bring the machine down but can cause referential integrity failure without a transaction based DB).

Poorly written DB apps do this. most of the time an app can easily have written in it simple steps to ensure it cant hose the DB. This certianly is not the fault of the db, whil there could be things added to prevent it, the DB is not there to make up for poor frontend software.

All that hardware you are using is great (essential) for 24x7 availability but overkill if you accept lower availability but still need consistent data.



No all that hardware is great if your data is IMPORTANT. I have nothing near high availability quality. If your server hardware did not cost from $4000.00 to $12,000.00 new then you don't have a server you have a toy. PERIOD. no if's and's or butts.

nothing should be expected to survive a power drop. Espically when it is very cheap to buy systems that can easily prevent this problem.

Yes, $15,000.00 each server is CHEAP for important data. 1 server, 1 enterprise class UPS, 1 LOCKABLE server rack. (this is assuming it is located in a air-conditioned room.

it sounds like you clients have more problems than database stability. I'm betting they dont even have Policy and Proceedure manuals for the IT department or even a reporting and checks and balances system. How about a server logbook?

I dont care if it's a 10,000 member IT force or a single IT person for the company... if data is important, the above is CRITICAL.

I suggest you reccomend to your clients someone you trust to replace their IT help and offer to write up a nice IT policy and proceedure manual to eliminate these problems... Hell you could make at least $3000.00 from a small client easily by doing that and then cookie-cutter with modifications to your other clients.

Re:I think PostgreSQL is more of a threat

[j3110]

MySQL 3.23.?? (I forgot the minor version since. approx 50)

Check table was just fine.
Select * was fine
select * where keyfield=keyvalue was not fine

alter table t type=myisam;
alter table t type=inno;
fixed the problem until it did it again.

This was a problematic server, and it could have been because of power loss if the keys aren't part of the transaction... I don't know how inno works all that well.

inno was running on it's own partions if that matters.

autocommit = 1 :) I wouldn't do it any other way... if I want a transaction that requires multiple statements, begin/commit was made for that.

I hand compiled this version of MySQL on a debian system, so my libraries/kernel may have been newer than inno.

This may just be a freak thing that only I have run in to, and only I may ever. inno was working for a long time until I put a high volume table into inno's hands as well.

Just be on a look out for anyone else who has the problem. I didn't know I was the only one. If I'm not, maybe you can find commonalities.

Re:I think PostgreSQL is more of a threat

[Heikki_Tuuri]

Hi!

It might be this bug which was reported recently and which I have now fixed in upcoming 3.23.57 and 4.0.12:

MySQL/InnoDB-3.23.56, March 17, 2003

* An outstanding bug: MySQL can erroneously return 'Empty set' if InnoDB estimates an index range size to 0 records though the range is not empty; MySQL also fails to do the next-key locking in the case of an empty index range.

I know that this bug can hit a table which has only one or a few rows. I am not sure if it can hit big tables. The reason for the bug was that MySQL assumed that the range cardinality estimate is always accurate.

Best regards,

Heikki Tuuri
Innobase Oy

Re:I think PostgreSQL is more of a threat

[j3110]

Yea!

Gotta love how quick you guys work on things... You have a great product. My only beef is really MySQL, you guys give developers, like me, who developed large applications using MySQL just to find that ACID compliance (particularly where corruption comes in) is more important than we thought an alternative. I'm still going to at least abstract the database and support other servers. I'll just use whichever database works better at the time :)

Thank you and your team mates for some great code. A lot of those people that love MySQL for no good reason will be using inno when they realize that there are a million things that can go wrong, and ACID compliance is the only way to ensure your database won't be corrupt. (and raid0... backups to remote locations)

Keep on inching MySQL/inno toward enterprise quality. Thanks again for your effort and code :)

20K reasons why it is

[Crashmarik]

I LOVE MYSQL.
Its well integrated into just about every web development system you can name.

With PHP or perl you can roll out a small interactive site in less than a day.

At least you can figure out how much it costs. I can't say how much customers love hearing about ORACLES price by the system you install it on system.

Its not one of Microsofts line of swiss chese products that have more holes than a typical sieve. Slammer worm anyone ?

Oh yeah MYSQL is a threat. Software that works and a revenue model that depends on satisfied customers.

Crash

postgres, schmostgres...

[twiggy]

I don't see Postgres getting much media coverage until the syntax stops sucking.

I'm sure that a lot of uber programmer types are going to say "if you can't figure it out, you're a jackass", etc etc... but the bottom line is too much easy stuff is a pain in the ass to do, syntactically, with Postgres... Even auto incrementing IDs in Postgres are annoyingly difficult compared to MySQL...

I am well aware of the cool stuff that postgres does, but a lot of that cool stuff is only needed by people with specialized purposes.. your average blog site or even e-commerce site doesn't really need nested queries and all that stuff, so why go through all the syntactical annoyances when doing a simpler site?

When pg gets easier to use, i'll make the transition...

Re:postgres, schmostgres...

[Jeffrey Baker]

If you can't figure it out, you are a jackass:

create table foo (bar int default nextval('my_sequence'))

That's hard? Give me a break. MySQL is so internally inconsistent that auto_increment is practically the only atom in the entire data definition syntax that uses the underscore! How about this bit of MySQL genius:

create table foo (bar int default 42 auto_increment primary key)

Which is the default: 42, or max(foo) + 1? The statement is internally inconsistent but MySQL allows it anyway. Nevermind the stupidity of requiring a unique index on the auto_increment column.

PostgreSQL has a number of operational problems -- vacuum, toast table indexes, and so forth -- but the SQL syntax is not one of those.

Re:postgres, schmostgres...

[optikSmoke]

Even auto incrementing IDs in Postgres are annoyingly difficult compared to MySQL...

Uhuh.......... CREATE TABLE foo (bar SERIAL PRIMARY KEY); (PostgreSQL)

CREATE TABLE foo (bar INT AUTO_INCREMENT PRIMARY KEY); (MySQL)

In both of the above tables, the bar column will behave pretty much the same way in each database. Yes, annoyingly difficult compared to MySQL because....... the syntax is different? Maybe if the first db you used was MySQL things seem "harder" because Postgres is a little different, but you'll see differences like that moving between any database.

In any case, even though I use Postgres, and prefer it over MySQL (which I have used extensively before), I am happy that opensource dbs are getting recognition out there.

Re:postgres, schmostgres...

[the eric conspiracy]

Even auto incrementing IDs in Postgres are annoyingly difficult compared to MySQL...

Nah, they actually make much more sense in Postgres than they do in MySQL because you can access them using a standard query rather than some bozo non-standard driver extension as in MySQL.

The problem with MySQL is that the lack of basic functionality like triggers, subselects, foreign keys makes it a total PITA except for the simplest applications. Sure you can write code that works around the implementation limitations, but WHY should I have to reinvent something over and over that should BE PART OF THE DATABASE?

You may think this stuff is esoteric, and not needed for the average blog or even e-commerce site, but that's baloney. FKEYS *ARE* needed for just about *ANY* database application except the most trivial - ie. an address book.

MySQL - forget it - it just isn't competitive with other free databases out there.

Re:postgres, schmostgres...

[GiMP]

foreign-keys are now in MySQL 4; however, they are quite a bit late.. as with all the other important features like views, subselects, etc.

Re:postgres, schmostgres...

[tzanger]

Maybe if the first db you used was MySQL things seem "harder" because Postgres is a little different,

Little different? How about it's doing it the right way since it is conforming to SQL95 specifications? The goddamned '*' autoincrement is but one more blight on MySQL's arse.

Re:postgres, schmostgres...

[tzanger]

ie. an address book.

And even then, you should be using LDAP.

Re:postgres, schmostgres...

[the eric conspiracy]

important features like views, subselects, etc.

Views, eh. MySQL isn't promising views until at least version 5. Triggers who knows when.

Re:postgres, schmostgres...

[TheLink]

Erm, haven't you checked the FAQ?

http://www.postgresql.org/docs/faqs/FAQ.html#4.1 5. 1

That FAQ item has been there for YEARS. I know coz I switched from MySQL to Postgresql.

Sure it's different from MySQL but doesn't look annoyingly difficult. Plus the design allows greater flexibility without much increase in complexity.

I can't recall where Postgresql's syntax sucks. Overall SQL's syntax sucks but that's not Postgresql's fault. For example: SQL inserts and updates have very different formats/syntax while they are doing a related thing, that seems so wrong to me.

Since you don't even read the FAQ, I suppose you're waiting for a "Matrix knowledge uploader".

"Tank, load the Postgresql Guru program", eyes flutter, etc etc.

MySQL is faster than Postgresql at loading lots of data into a table. Postgresql seems rather slow at that even compared to other databases - not sure why.

MySQL has replication and win32 support. Both of these for Postgresql are not fully baked at the moment. The latter is not really a problem to me, heck sometimes that's a feature when talking to some bosses (No sorry boss it doesn't run on Win98, but we can run it on Linux/FreeBSD and that's free too).

MySQL works with XXXX and XXXX doesn't support Postgresql.

Re:postgres, schmostgres...

[Ragica]

Or just simply:

create table foo (bar serial);

Add "primary key" after the "serial" if desired. Done.

Re:If MySQL was just a bit more user-friendly...

[drudd]

Have you ever looked at MySQL's online documentation? It's wonderful...

Fully indexed, with user comments... I often find new techniques while searching for something completely unrelated. I think the great documentation is one of the reasons why MySQL has taken off, it's just so easy to learn.

Doug

Re:If MySQL was just a bit more user-friendly...

[LiENUS]

sure you can get support from mysql
http://www.mysql.com/consulting/index.html
whats that?

SAP DB

[wuchang]

There's also a lot of interest in SAP DB

Re:SAP DB

[exhilaration]

But nobody seems to care. I'm a fan of PostgreSQL, but I'd love to read some articles comparing SAP DB to the tools we're already familiar with.

SAP DB is a true, tested enterprise database - did anybody out there start using it when SAP open sourced it? Anybody??

Re:SAP DB

[it0]

Where did you get the Idea that SAP DB is enterprise tested? SAP is installed 99% of the time on oracle.

Re:SAP DB

[Dion]

Yes, SAP DB is very cool, it is as "Enterprise" as any other database on the planet, it doesn't have some of the bloaty features that Oracle or MSSQL have like full text indexing or function based indexes, but it is a real database and it implements all the database functionality in a very robust manner.

A datapoint is that SAP use it to run SAP R/3 on, that system has over 16000 tables and 2GB of data in a newly installed database.

Performance wise it's nice as well, SAP said that in their benchmarks all the databases they support (including DB2, MSSQL, Oracle and SAP DB) the difference between the fastest and slowest was less than 5%.

Here we use it to back a webapplication that has 134 tables in the database and a little over 120MB in the database and it's quite happy.

We have crashed machines with SAP DB many times, either due to bad drivers, crazy software or poweroutages and we have never lost any data that was committed.

SAP DB doesn't need to use the filesystem like those "little" databases, like the other big DBMSs it can use raw devices directly and it will distribute data on the devices for better performance, it can even execute one query on more than one cpu for better performance.

SAP DB is simply better as well as GPL.

I wonder who mysql steals marketshare from?

[Billly Gates]

It seems oracle is going nowhere in terms of its core market. I am hoping its eating middle end and low end databases like ms-sql server. Access in the low end isn't going anywhere because of its gui and development tools.

I wonder about Sybase and Paradox which seem to be mid to high end of the market which Microsoft really hurt and now so its mysql.

I tried out mysql and its ok but postgreSQL is alot better for a RDBM. Its no wonder that RedHat picked postgreSQL for its database product. In Asia the situation is opposite of the west and all the technical books are for postgreSQL.

I just find it hard to believe its eating Oracle's or IBM's core markets. Mysql is a simple bicycle vs a high end car in comparison.

Re:I wonder who mysql steals marketshare from?

[RevAaron]

I was having a discussion with a coworker a month or so back about the differences between PostgreSQL and MySQL. I was under the impression that Postgres was a lot more like a real enterprise db than MySQL and that transactions was a huge part of that. Looked it up, MySQL recieved those pretty recently... other than that, what does PostgreSQL still have on MySQL? This guy is pretty uninformed, yet very confidant in his assertion that "MyEthQueElll ith ath good ath Oraclee!!" and it'd be nice to have the material for enlightenment. :)

Re:I wonder who mysql steals marketshare from?

[Billly Gates]

Sure.

Look here and here. Both of these websites mention what postgreSQL (and Oracle)offers that mysql is lacking as well as how to migrate to PostgreSQL. Keep in mind I am not a database administrator or do I consider myself a sql guru. I only use them to write web enabled apps as a hobby and not in a corporate environment.

However it was rumoured that postgreSQL lacked real backup tools to fix a corrupt database. I believe this might of been fixed but was an issue 3 years ago. This is the only downside I see. Both Mysql and Oracle have tools to fix such a problem. Maybe someone who is reading this who is more familiar with administering databases can comment on this.

Re:I wonder who mysql steals marketshare from?

[TobiasSodergren]

I asked something like this, a while ago. There were some good answers attached to it, some which might be of interest.

Re:I wonder who mysql steals marketshare from?

[tzanger]

other than that, what does PostgreSQL still have on MySQL?

Uh... Performance? IIRC MySQL's benchmarks are horrible for anything other than simple selects. Even its network performance is weak compared to Postgres. If I can find the benchmarks I'll reply here, but IIRC it was even in a previous /. story. :-)

Re:I wonder who mysql steals marketshare from?

[Billly Gates]

Also look here.

This shows how much postgreSQL can really scale.

Re:I wonder who mysql steals marketshare from?

[Hrunting]

Look here [webtechniques.com] and here [sitepoint.com]. Both of these websites mention what postgreSQL (and Oracle)offers that mysql is lacking as well as how to migrate to PostgreSQL.

The first article was written in September 2001. The second article was written in October 2001. The person who replied to your post cited an article from 2000, almost three years ago. The PostgreSQL vs. MySQL argument would be a whole lot more interesting if the articles cited were actually relevant to newer versions of both databases. It would also be great if they were more than just, "Hey, look, I got my inefficient bulletin board working a little better under database XYZ."

The best database is the one that has the features you need, the performance you desire, at a price you're willing to pay.

Re:I wonder who mysql steals marketshare from?

[RevAaron]

That was one of the problems I ran into- most of the comparisons were a year old or more, and a fair amount has changed since then.

I agree, your needs should dictate what you use, but I think the needs has been specified- something in the domain of "enterprise relational databases." Not the hardest description, no, but something in the league of Oracle, DB2 and Sybase. For PostgreSQL to be "better" than MySQL for this area, why? Such is my questions for those who were not able to figure that out.

Performance seems to be a big part of the equation, and PostgreSQL has more of it?

Re:I wonder who mysql steals marketshare from?

[sql*kitten]

It seems oracle is going nowhere in terms of its core market. I am hoping its eating middle end and low end databases like ms-sql server. Access in the low end isn't going anywhere because of its gui and development tools.

MySQL is "stealing" marketshare from CSV and grep. Anyone who could use MySQL but instead uses Oracle has wasted their money! The two products aren't even comparable.

Re:I wonder who mysql steals marketshare from?

[rsax]

Access in the low end isn't going anywhere because of its gui and development tools.

Well, if you really need Access' GUI and development tools then you can just use it as a frontend and have PostgreSQL be the backend using ODBC.

Re:If MySQL was just a bit more user-friendly...

[Thorizdin]

umm perhaps you should take a look at the actual support options Here is their features chart I doubt you can find another DB company whose support options include escalation to the CTO when needed. BTW CTO in this case is Monty himself. Pricing for support is here

That depends...

[dasmegabyte]

A lot of the viability of any product is based on who is selling it as part of an end to end solution. More and more developers are doing this with MySQL, but most of these developers are doing it for relatively low end applications. The "high class" consutlants and developers will be using DB2 and MS SQL forever, because a) they're told to do so by the people who can fire them b) they're used to it, and used to touting its glory c) they have a ton of tools for it.

Furthermore, there are some applications that just don't make any sense to switch. An example is government databases. I'm working right now with a state government database written on top of Sybase, and i don't think it's ever going to move off of Sybase unless the company tanks. There's actually three pages of (somewhat unfounded) explanations as to why it can't be ported to MS SQL. Mostly bullshit about WACOM SQL being incompatible with Transact (which begs the question, why not just use Transact in the first place when MS' and Sybase' version are about 80% similar). Can you imagine the developers, who have big enough egos to include three pages of MS SQL Server bashing in their docs, redoing their whole bloated app just so it can run on a free environment? Lord no! Not to mention the cost to taxpayers, who have already footed massive bonds to pay the usually high up front costs for software. Think they're going to pay a hundred k for some developers to rewrite everything in a free environment when they could just pay a few thousand for a Sybase license?

Do I think that truly open minded (some would say wise) development houses looking to cut costs on new systems are going to go MySQL? Absolutely. But there'll always be a place for the behemoth server app, not because it's better, but because it's PERCEIVED as better.

Re:That depends...

[jfpoole]

There's actually three pages of (somewhat unfounded) explanations as to why it can't be ported to MS SQL. Mostly bullshit about WACOM SQL being incompatible with Transact (which begs the question, why not just use Transact in the first place when MS' and Sybase' version are about 80% similar).

Depending on which Sybase database they're running, it might not be unfounded. There are a couple of Sybase database servers out there. One is Adaptive Server Enterprise (ASE), which is based off the same codebase as Microsoft SQL Server. Another is Adaptive Server Anywhere (ASA, formerly SQL Anywhere, formerly Watcom SQL), which is a different codebase entirely. If the developers are talking about Watcom SQL, then I suspect they're running ASA, not ASE, in which case porting an application from ASA to SQL Server might be non-trivial (I've no idea, since I've never tried it).

There's also the fact that ASA is fairly cheap as far as database servers go, so there might not be much incentive to port these applications to MySQL.

(Disclosure: I'm a code monkey for Sybase.)

Re:That depends...

[darnok]

> The "high class" consutlants and developers will
> be using DB2 and MS SQL forever, because a)
> they're told to do so by the people who can fire
> them b) they're used to it, and used to touting
> its glory c) they have a ton of tools for it.

On the contrary, I get paid a reasonably high consulting rate (well, it's high for the current climate...) because I save people from having to fork out for DB2, Oracle, etc.

Think about it: if you can front up to the decision makers in an organization and demonstrate conclusively how you can save them large slabs of money, then you're in a good position to ask for a cut of what they've saved. Tell me one IT decision maker these days who won't listen to someone credible who offers to save them hundreds of thousands or millions of dollars in software licence fees...

Even at the lower levels, IT departments are being asked to shed staff, and if you can tell them they can cut costs elsewhere and maybe retain their existing staff as a result, you'll generally get someone prepared to listen.

Re:That depends...

[Quintin Stone]

I'm at my third state project, and my state uses DB2 for EVERYTHING. God, I hate it. It's such trash. Scales well, I suppose, but I just hate using it.

Re:That depends...

[dasmegabyte]

Yes, but when the total cost of switching to an OSS solution (including updating all your current apps, importing and maybe transforming your current data, retraining your dba, redeploying the app to your clients) is greater than just shelling out for upgrades/new licenses on the old and presumably reliable for-pay database, smart money should stick with the old stuff.

And when your apps are costly custom work as it is, you've got to deal with the possibility of the cost being equal to the cost of a new application.

In the end, if you've got a single repository situation, you're not paying all that much per SQL license anyway. It's not quite the same as deploying software to every client machine. The savings aren't necessarily scalable. And certainly there's still some peace of mind in going with a big, reliable, and most importantly ACCOUNTABLE solution: if the server fails, you can always sue.

This isn't to bash OSS DBs. I love postgres and would love to dump all my garbage MS SQL stuff for it. But compared with other savings possibilities from OSS (like switching OSs to Linux, web servers to Apache, or email packages to qmail), there are lots of examples where switching to an OSS DB from an established DB system wouldn't be cost effective.

In short: if it ain't broke, don't break it just to prove a point.

Redundancy

[LegendLength]

One feature which I think would help it greatly in the corporate world is 'automatic' redundancy. Where, if you have a database shared by multiple servers, pulling a server out doesn't mean that part of the database is lost.

I know Oracle can do this but I believe SQL server cannot.

Re:Redundancy

[Pseudonym]

Oracle doesn't have "automatic" redundancy, or at least not in the sense you mean. You still need to code your application to handle servers which appear and disappear. While it may be "automatic", it's not "transparent".

Re:Redundancy

[silas_moeckel]

Actualy running Paralell server with Oracle's shim you can pretty mcuh remove hardware and not notice it; at least this workded very well on an SP cluster running parallel server under AIX we lost a node and didn't notice anything on the things connected to them.

Re:Redundancy

[Dave2 Wickham]

Re:Redundancy (Score:1, Redundant)

Hehehe...

Threat?

[gmuslera]

It could be benefical for big databases. With free and widely available databases more and more applications will rely in it, and for those applications that need to grow more than the actual database can, then there is where the big databases come. After all, all are SQL based, is easier to migrate an application from mysql/postgressql/msql/interbase to oracle/redbrick than, well, a non sql database to a sql one.

In a job I had to migrate an application done originally in clipper to web/sql/etc, and choosed mysql because I thinked that it will be enough, but if not, the migration to a new sql server will be a lot faster and less complex than the first one.

Re:If MySQL was just a bit more user-friendly...

[gl4ss]

all ms software is 'at your own risk' software.

very much so more because you can't even do anything with them if ms decides so that you don't deserve it to work anymore. mysql is pretty much easy enough to install for somebody who wants to have alternative to ms-sql server, however sometimes people don't for very questionable reasons even want to look for change(and i'm not saying mysql would be the only best option either, just that most windoze users can even install it, mysql's website does offer very good docs btw, and i thought support is what where the money mentioned comes in? and the point of the article is that it is __not__ a fringe, experimental db anymore, instead it is very widely deployed).

Re:Ethical obligation?

[ccady]

Sorry.

The GPL, one of the licenses under which MySQL is distributed, states that if you re-distribute it, you are also required to share the changed source code.

My complaint was that the article was imprecise. If a company changes, but does not re-distribute MySQL, they are under no obligation at all, ethical or legal. If they re-distribute it then they are under a legal obligation to share their changes to anyone who uses it (not just MySQL AB).

"Ethically Obliged"?

[Landaras]

From the article:

Anybody can download the product for free and use it for whatever they want, but in so doing they become ethically obliged to share any modifications with the company.

The GPL does not merely give you an ethical obligation to share your modifications with anyone you distribute them to. It gives you a legal obligation. Until shown otherwise by a court, the GPL is legally binding. As such, stating that the (presumably only) obligation that someone modifying the code has in an ethical one furthers the outdated notion that all pieces of Open Source Software are amateur projects that are only held together by people who choose to donate their time for whatever higher reason. Not that there is anything wrong with volunteering your skills, but there are major businesses investing time and money in OSS.

From a business standpoint, OSS is legitimate. It would be nice if CNN reported it that way.

Note: I contacted CNN.com regarding this when they first posted the article. Predictably, I have not yet received a response.

Re:"Ethically Obliged"?

[RevAaron]

I don't think the writer was confused in that the GPL is legally binding rather than ethically binding. That is, the GPL only legally requires you to redistribute your code if you pass out/sell the binary, not if you make the changes for your in-house setup. However, if I am a business using a heavily modified version of MySQL, adding tons of great features that make it a real player with real enterprise databases- but not sharing or selling the binary, there is still an ethical obligation- not a legal one, pressure from the community at large to share your changes. You see it all the time in the Linux community in especial.

That is how I read that statement, and from that standpoint, the author is correct.

Re:"Ethically Obliged"?

[Landaras]

Agreed: There is no obligation, ethical or otherwise, to "share any modifications with the company" (in this case MySQL AB) if you are not directly distributing those modifications to them. However, what I quoted was from the writer of the article, not MySQL AB's CEO.

Re:"Ethically Obliged"?

[Tony-A]

use it for whatever they want, but in so doing they become ethically obliged to share any modifications with the company.
There is a legal requirement to make any modification available to any idiot that would buy TonySQL. (However I'd be a bigger idiot to try to keep up with the original;)
I think what CNN is trying to get at, is that if I've a few things that are actually useful (to me at least), it is strongly in my own self-interest to go to a lot of extra effort to get the stuff into the main line. Calling it "ethically obliged" is a misnomer, but it is an effective way of viewing the situation.

Re:"Ethically Obliged"?

[frostman]

Until shown otherwise by a court, the GPL is legally binding.

I think it's more like "until shown one way or another by a court, the GPL is probably legally binding, so unless you can afford to lose the fight it would be unwise to violate it."

The GPL is a contract which I implicitly accept, and many legal systems give broad rights to the creators of intellectual property, strengthening the "accept" part of it.

Also of course GPL is only one of many good ways to license your open source software.

Otherwise, I agree with you.

Re:"Ethically Obliged"?

[sootman]

Anybody can download the product for free and use it for whatever they want, but in so doing they become ethically obliged to share any modifications with the company.

ONLY IF THEY WANT TO REDISTRIBUTE THEIR CHANGES!!! &*%$#@*#&$%, do we have to go through this every time? Once more for the cheap seats: You can download GPL software, do whatever the &*$#% you want to with it, and NEVER TELL ANYONE, if you keep your modified version in-house. OK? Are we all clear on that?

Re:Ethical obligation?

[smart.id]

RTFA. This article is from FORBES.COM. It was not directed towards the users of Slashdot, it was for the general public. The journalists write things like this so it is easier for Joe Sixpack to read. They could include an analysis of the GPL, it's strengths and weakesses, or they could give a quick blurb about it.

Reason why PostgreSQL or mSQL weren't mentioned

[ciurana]

From the posting:

Of course, there's no mention of PostgreSQL or mSQL, but I guess that's typical.

This article has all the signs of being the effort of MySQL's PR firm. Nothing wrong with that; they didn't mention PostgreSQL or other OSS databases because their desired outcome is to increase awareness of MySQL, not the others.

Cheers!

E

Re:Reason why PostgreSQL or mSQL weren't mentioned

[leviramsey]

That is the way that things tend to work in the magazine world. In this case, since MySQL is not that big a fish, Fortune's editor tells some junior writer, "Forbes and BusinessWeek have run articles on open source software in business. Do something on open source databases." The junior writer then looks at the figures, sees that MySQL is the most popular of the OSS db's, sees that there's an actual company behind them, and calls up MySQL AB's press office. Said PR firm basically sends him an outline, with relevant quotes from luminaries, and junior writer bangs out a story from that outline.

Moral of the story: magazine writers are lazy.

Cheap, fast and easy.

[Hamstaus]

MySQL is a phenomenal product, in terms of just how much a small to medium size business can accomplish with it, for so little cost.

Having to use a data-storage solution like Oracle is simply unfeasible for anyone but large companies. I've been using MySQL for 3 years to build web applications, and I've never had a crash or corrupted data. The only problems I ever ran into was when one of my systems had a table get to 2GB on the 2.2 kernel, but that wasn't MySQL's fault ;)

With the inclusion of InnoDB, MySQL definitely becomes a threat. The main problems I've run into with MySQL is backing up/restoring without locking up the whole system (table-level locking). InnoDB of course removes this!

I see no reason to use Oracle over MySQL for anything but the largest system. Then again, why even that? Doesn't Slashdot run on InnoDB...?

Re:Cheap, fast and easy.

[ckaminski]

Yup. And how many times do you see crosslinked comments posted to the wrong stories?
Arguably this could be a bug in Slashcode, but using Slashdot as the benchmark for anything has it's problems.

Re:Cheap, fast and easy.

[Hamstaus]

I doubt if the crosslinked comments have anything to do with their database system. I find it hard to believe that it would simply mix up comment and story ids, and not screw anything else up. That's likely a product of Slashdot moving their stories around.

Re:Cheap, fast and easy.

[Hamstaus]

I'm afraid I have to disagree with you about that being "a lot worse". The hot backup tool costs $440 US (400 Euros). This is a piddling sum compared to not having any downtime. If some of the systems I worked with were down for the time it took to restore the data to master/slave scenarios, or take a snapshot, or move tables around or whatever, we could easily lose that much money in a single hour of downtime. Consider this: your main database server needs to be replaced, and has several 2GB+ tables, and you need a snapshot to set up a new master/slave somewhere else. Locking tables is not an option - it would take down the system. An hour of downtime could easily cost $440 US... Considering the other option is buying Oracle for thousands of dollars, it's not like you're breaking the bank.

Re:Cheap, fast and easy.

[bovinewasteproduct]

The hot backup tool costs $440 US (400 Euros).

Considering the other option is buying Oracle for thousands of dollars, it's not like you're breaking the bank.

Why do I have to goto Oracle? What about PostgreSQL and Firebird?

BWP

web databases != database market

[Misha]

you probably won't find too many databases on the 'net that need the kind of performance some commercial brands give. so i wouldn't say the drastic change is coming, unless companies start putting their payroll records for the web to see.

our company actually puts mysql onto websites, but no client comes (at least for us) and says 'can you replace my blah-blah db version blah point blah with mysql'. we usually put mysql as a replacement for product databases, forums, etc. which previously were stored in text files or worse. and we usually do this for clients who simply can't afford anything and haven't invested into updating their site in 1-2 years. if they can afford it, they usually already know what they want, and it usually doesn't come free in a cvs snapshot.

Re:web databases != database market

[tupps]

But if you said to the client that putting a DB on the website would cost the thousands that it does for a MS SQL Server Web License then I am sure you would have more people asking for it.

Re:web databases != database market

[g4dget]

you probably won't find too many databases on the 'net that need the kind of performance some commercial brands give.

Quite true: open source databases are often much faster than their commercial counterparts. Of course, there is a reason commercial databases are such a dog: it's all the features they support and all the transactional and consistency guarantees they are trying to make. The question is: do people need all that? And the answer may well turn out to be no for most database applications.

Re:web databases != database market

[Misha]

actually we do.

Re:web databases != database market

[Unordained]

firebird is open-source, does transactional isolation, etc. like the big guys. it's fast, small, and free. i've heard from people using postgresql that the two are quite similar, feature-wise. firebird (as interbase) has been around for a -long- time, and has pretty much always had transactional isolation (it's not an add-on after-thought like, say, with mysql, or less so, with postgresql.) and did i mention that the original designers and coders now work on firebird, not interbase?

and it -is- necessary to have that isolation. i dunno where our application would be without it. it's not a web-app though, and that makes a difference -- we have transactions open for minutes, maybe hours at a time, with individual client machines doing whatever they need to do.

what it doesn't have -- stuff like xml (what for?) or tables as domains, or full-text indexing (note: even in oracle, the full-text indexing wasn't an oracle-built package, as i recall.) i think its support for unicode is improving, but it wasn't built with that in mind. (unicode on a pdp-11?) they're switching the code base to c++, taking -out- support for platforms nobody cares about. (helps reduce the size.)

i agree you can get more speed out of something without transactional isolation -- but i hope you never need it. when it rains, it pours ... and i've run into the need for transactions even in web-apps. every millisecond counts, not speed-wise, but logic-wise.

Re:web databases != database market

[g4dget]

i agree you can get more speed out of something without transactional isolation -- but i hope you never need it. when it rains, it pours ... and i've run into the need for transactions even in web-apps. every millisecond counts, not speed-wise, but logic-wise.

Of course, transactions are needed in many applications--you don't want to sell the same hotel room twice. However, the database is only one of many places where transactions can be implemented, and it is often not the best place. Of course, most people trained in commercial computing won't even consider other design possibilities.

Re:web databases != database market

[einhverfr]

I do think that these are different but web databases are a market segment, and one that is growing.

So I think that we are seeing a change in the market because databases are becoming more ubiqutous in part due to MySQL.

But that does not mean that they are stealing Oracle, et. al's market. Instead they are just finding a use for a certain sort of project.

Re:web databases != database market

[Unordained]

i'm completely willing to consider other design possibilities. i do it all the time. but in this case, i'd say the end-to-end argument doesn't go very far. i wouldn't, in the case of transactions, want an outside system doing the following:
- deciding whether or not my commit succeeded (safe-write to disk)
- deciding which tuples i can see (and in what state)
- deciding who gets to write to a tuple (when the database, and its decision as to which serial request to consider next, has the last word on that one) ...

i can do simple locking elsewhere -- for example, we have functionality that doesn't -obviously- get protected by database transaction isolation: when it's mostly inserts, and none of the relation-constraints have anything to say about it ... you -can- resort to, for example, having a 'lock' table containing tuples for each operation you want to maintain locks on, and then lock the rows (for update) as you start your operation. you could -also- do it in an outside system. so long as your local logic respects the results, you're fine. or you could guarantee you won't have trouble, and actually put all the operations on one transaction, handled by one agent (the database server app.)

on a similar note: i agree individual applications -might- be able to manage (virtual) memory better than the OS. but would i recommend it? no. there's a limited resource, with many agents vying for it. it makes sense for a central agent to manage the resource. an OS usually manages files, memory, ports, and other resources to avoid conflicts; i expect a database management system to do the same for my data. i like modular designs. i really do. but that's just one chunk i'd rather not out-source.

Re:Ethical obligation?

[RevAaron]

Read what you put in bold again. This journalist said that they become ethically obliged- not legally by the GPL. And largely, that is the case. It would be nice if the writer could have said more than just that, but how could it have been said better in one sentence?

If you don't think that businesses (to a lesser extent, individuals) have an ethical obligation imposed by the community to share modifications made to an OSS project, you haven't been reading Slashdot or other forums of Free/OS software discussion. Hell, half of the "Ask Slashdots" where someone is asking about some whether application that does x, y, and z exists for Linux, there is always a big hunk of replies stating "this is OPEN sourze d00d write it yourself and share it with the world. otherwise, you don't deserve to use OSS- even if this app already existed." Yes, it's a stupid attitude (IMHO), but quite real in the community.

Tools

[Elektroschock]

GPL Admin tools like TOra 8very good, targeted at Oracle) and some KDE Tools like Kexi will prove that MySQL ist a professional database. However, more important is easy administration. Imagine MYSQL without PHP!

Who Wants To Be A Millionaire?

[jfisherwa]

Put a wrapper/installer around MySQL or PostgreSQL that lets you import a SQL Server database dump -- including stored procedures.

Re:Who Wants To Be A Millionaire?

[rycamor]

It's going to take more than a wrapper to get stored procedures into MySQL (lol).

But that is an excellent idea for PostgreSQL.

Re:Who Wants To Be A Millionaire?

[pbur]

someone is working on something like that...but for coverting from almost all databases to each other (ie SQL Server to Oracle, SQL Server to DB2...etc)

It looks like Postgresql conversions are "Coming Soon"...here's hoping they can do it.

http://www.realsoftstudio.com/

Re:Who Wants To Be A Millionaire?

[airgee]

they will use table functions.
Available in PostgreSQL 7.3.x (latest stable version).

Re:If MySQL was just a bit more user-friendly...

[tuffy]

Fully indexed, with user comments... I often find new techniques while searching for something completely unrelated. I think the great documentation is one of the reasons why MySQL has taken off, it's just so easy to learn.

Furthermore, look at the dead-tree documentation available for each. Not only are there more MySQL books available than PostgreSQL ones, but the MySQL ones tend to be larger. For example, the New Riders MySQL book is over 400 pages longer than the PostgreSQL one - and their MySQL title isn't exactly brimming with filler. I'm convincesd PostgreSQL adoption will lag behind MySQL until more/better docs show up so people can learn it and its capabilities better.

Re:If MySQL was just a bit more user-friendly...

[harvardian]

Your comment is spoken from the point of view of somebody who runs a website with a small to moderately sized, simple database.

So that's great, you blog well in PHP, but you utterly overstate the power of MySQL. MySQL will not even approach taking on the big database apps until they have a) subqueries and b) true ACID support. Many complex websites have stored procedures with subqueries in them, and it's logistically impossible for these sites to migrate to MySQL, since it would involve rewriting anything with a subquery (if it's possible at all without nasty temp table hacks).

And not having ACID (atomicity, consistency, isolation, and durability, btw) support? No enterprise-level website, especially one that does eCommerce, would ever think of running a database without proper transactions.

So, yeah, MySQL is great at handling small- to mid-size loads, but once you get to high complexity or you need transactions, it's over. Maybe if they include this stuff in version 4 and it survives a couple of years of testing, THEN we might see some significant migration, but it will not happen until then.

MySQL vs "bigwigs"

[mabu]

Let's see, I can pay $14,000+ for an Oracle license, which involves a deliberately convoluted array of hoops to jump through, "education points", maintenance contracts, and product module/pricing options that would make most long distance plans seem trivial. If I need support I have a plethora of options with cool names like "bronze", "silver" and "gold" which in effect give me a varying scale of hours or days in which I can wait to have issues resolved.

Or I can pay $300 for a MySQL commercial license and if I have any problems or additional functionality there are thousands of web sites with tons of free advice and code. And MySQL will blow the doors off of Oracle and other databases in terms of raw speed. Advanced options like data warehousing, replication, fancy triggers, rollbacks and other systems can be sometimes better-integrated on a hardware/os level so it's arguable as to whether MySQL's lack of some features is even a negative. You get more with less all around with MySQL: better support, more rapid bug/security patches, less hardware requirements, and more people in the online community who are willing to help without charging you by the hour because of their pretentious DBA title.

Tough choice...

Re:MySQL vs "bigwigs"

[LoztInSpace]

Rollback is not an advanced option. It's essensial for any non read-only database. When you have a typical OO or other modular system you often ask a particular function to do something without it being aware of the wider context. Something in that wider context may well decide something is wrong. Without rollback (and commit) life becomes a royal pain in the arse and/or your data becomes rat shit.
Replication is useful for keeping things going 24/7 and allowing intensive off-line reports etc.
Lets not confuse "advanced" or "bloated" features with "features I don't use".

Re:MySQL vs "bigwigs"

[jonr]

What idiot marked this as "Insightful"? Let me make a bullet list:

• You can find lots of documents online at oracle, but I guess you didnt care
• Speed, MySQL is speedy, because it doesn't have to do anything
• Again, no refirental integry (OOps, I didn't mean to change that foreign key, now I don't know what it pointed too, my database is corrupt! AAARG!)
• Rollbacks are not fancy stuff, they are essential to a real database
• Hardware levels? Ok, go back to your parents basement

With Oracle you can choose: speed, security.
Mysql: Speed. Only.
Mysql is like a dragster, fast but no control.
Next time you want to start karmawhoring, at least pretend that you know what you are talking about.

Re:MySQL vs "bigwigs"

[mabu]

You underestimate the significance DESIGN plays in the ultimate performance of a project. This is a classic example of the new breed of computer programmers who rely more on the corporate feed trough of tools, complexity and ongoing jacked-up hourly support rates, than the inherent resourcefulness of humanity.

It's particularly bad with Oracle, because once I was on the DBA track with them, I realized it made the legal professions' "Bar" look shallow. Oracle DBAs are obsessive in their quest to make other people feel their technology is almost magical and incapable of mere non-Oraclephobes mundane understanding. Thus the legacy of the Oracle DBA getting uber rates continues.

Don't even get me started about Oracle. The problem with Oracle (and I know. I'm an architect of one of the most successful, complicated and highest-trafficked web sites on the net using their system) is the convolution of their support system. I don't know what kind of "Oracle" work you were doing. Maybe it was remedial and you could find a simple FAQ to tell you how to run a shell script, but when you get into the high end game on Oracle, lots of bugs and performance issues surface that you have to basically prove to them you're aware of before they'll admit it.

Oracle: security ?

Please.

Yeah it's offtopic

[Bruha]

But I wrote this story to /. over 12 hours ago and I feel it was a little more than interesting. This guys story is nothing but cut and pastes of 3 sentences in the story. I would think that the relative success of MySql would be a bigger thing on /. since that's exactly what you guys use as your backend.

Come on guys first you cant keep from posting sometimes 2 to 3 duplicates of a story now you have to either makeup stories and users and post them instead of what people are submitting. I run a news site and I post every tidbit of what my readers send about the topics we cover without having to resort to what /. is starting to become a pile of mush.

He mentions none of the interesting things I proposed such as do you think some .com problems are due to the insane prices of NON OSS material such as Oracle or MS software.

On another note how many jobs do state and local governments cut just to pay non OSS license fees? For the price of putting MS office in each cop's car laptop how many more officers could you have defending the public from crooks and terrorists.

Re:Yeah it's offtopic

[Nix0n]

Not as bad as being Moderated Troll for saying almost exactly the same thing( first, I might add )as something Moderated +5.

Welcome to Slashdot, I sure hope you didn't expect fairness.

Mindshare

[arvindn]

MySQl being more popular than Postgres has a lot to do with mindshare than product quality. Take me, for instance. I set up apache for the first time a month ago, and I wanted a db server for some things. I had heard of both MySQL and Postgres, but I had been bombarded with the words "LAMP" and MySQL guide/tutorial/howto so many times in the past that my first thought was to give MySQL a try. I found it was already installed on my machine, had lots of documentation, and had no learning curve - no complaints at all. So, Postgres didn't even get a fair consideration from me. Of course, you might say that newbies and students like me don't count, but keep in mind that I might become a database admin some day, at t which point I would have a lot more experience with MySQL than Postgres...

Re:Mindshare

[PrimeNumber]

Another thing IMHO helping MySQL more than PostgreSQL in terms of mindshare, is the fact (at least for PHP 4.0+ I know for a fact 4.2.3) is that the MySQL support is loaded/compiled by default, but if you choose to use Postgres you have to include the --with-pgsql option (assumes location is /usr/local/pgsql) when you ./configure.

Personally it doesn't matter to me, I have both on my box (to each their own, etc). But since a good number of MySQL installs are used right along with Apache and PHP, it might actually be more of a convenience factor to explain why people might use MySQL more often than PostgreSQL.

Re:Mindshare

[rtaylor]

With all due respect, you're not going to make a very good database admin if your experience is limited to MySQL.

Much as I wouldn't hire someone who's sole unix experience was with Linux (for any position other than Junior anyway). You simply learn a slew more tricks of the trade when your experience is diversified. When making a decision, you can usually back it up with a decent reason rather than simply "It's what I'm used to".

Nothing against Linux or MySQL. I've said the same to people who have solely used Oracle on Solaris.

Re:If MySQL was just a bit more user-friendly...

[mitcharoni]

MySQL has the power (pretty much) to replace MS-SQL Server.

OMG, you didn't really mean that, did you? Oh, that's so cute...

MySQL is barely ACID compliant, doesn't support triggers or stored procs or views (just for starters), and you say it has the power to replace MSSQL Server?!?!? For goodness sakes, MySQL just NOW has a shared SQL area (query cache). You gotta crawl before you walk, and you gotta walk before you can run with the big boys. MySQL is a very capable database in its own right, but it's still in it's infantcy.

With some of the best replication and datawarehousing functionality on the market and consistently a price/performance leader, I don't think MSSQL Server is going anywhere anytime soon.

mySQL also runs under under Windows ...

[MarkMac]

Furthermore, neither PostgreSQL or mSQL have decent ports (e.g. easy to setup) on Windows (PostgreSQL requires installing Cygwin first)- mySQL does. This allows mySQL to compete with the "bigwigs" in this market (most particularly SQL Server). mySQL also works quite well on Solaris. Although the Linux server market share is increasing, especially for web servers, it still is not yet a major platform for commercial database software.

Re:mySQL also runs under under Windows ...

[mla_anderson]

The latest Windows version of Postgres still requires Cygwin but it comes in the .msi file and everything just works on install. It's very easy. Of course I don't know why anyone would want to put a production dB on a Windows machine, but for development it sure comes in handy.

Re:Ethical obligation?

[Micah]

Agreed that that statement isn't 100% accurate, but more to the point....

why are they putting this FUD in the article anyway? How many business users of MySQL are going to modify it? Probably very, very few. You just install it and it works. If you need something MySQL can't do, you'll use PostgreSQL or pay for one of the big boys.

This type of FUD might just scare away legitimate potential users who have NOTHING to worry about.

transactions

[nemeosis]

MySQL is good for certain applications, where you only read data, and don't write too much data. This works out especially well for most web sites, since they serve information, but doesn't necessarily allow too much information to be posted by the user.

Lots of message boards on the web use MySQL as their database, because even though people are uploading comments, the amount of data that they upload isn't all that much. Slashdot for example, a popular discussion could prompt 500 messages to be posted in 15 minutes, but still, that's not that much information.

The key word here is transactions, the constant reading/writing, downloading/uploading of information on a massive scale, where each occurence is audited. And I think that's where MySQL has its weakness. PostgreSQL is supposed to be a bit slower, but it takes transactions into account. Red Hat's database software runs on the PostgreSQL engine specifically because of this.

Banking and finance applications require this accountability, because it's just that important. Websites don't need that accountability and overhead, which is why MySQL shines for web servers.

Re:transactions

[Tony-A]

MySQL is good for reading and writing. Just not the same thing at the same time.
MySQL is not good for reading and writing the same table at the same time.
MySQL is lousy at handling writing and slow readers, enough so that I'd put slow readers on a separate slaved system.
If everything serializes nicely in the allotted time, MySQL will perform beautifully. When it doesn't, you get a cliff-edge like thrashing in the old time-sharing systems. When it doesn't, you'll find you need a much bigger and faster system.
About transactions. They are logically required when you have two or more things to update that must both succeed or both fail. The downside of transactions is that your transaction can fail for reasons outside of what you're doing. If you transfer funds from checking to savings at the same time that interest is updated on your savings account, One of the transactions had better fail. Now life gets interesting if the bank cannot update interest payments because customers are messing with their accounts.

Is Oracle doomed?

[randall_burns]

A lot of folks have focused on Microsoft as the company that ought to be put out of business by Open Source development. I would suggest though that some other companies are in a lot more immediate danger.

BEA is already competing against an Open Source product-and loosing the battle. Oracle could be a fairly early casualty. A big chunk of the appeal of Oracle rests on its ability to compete in industry standard benchmarks. Open Source products have historically tended lang behind in having "glitzy" interfaces-but have tended to excel in reliabilitly and performance. It is clearly a logical development that in the next few years, the Oracle database will find itself replaced by an Open Source Database-this will be a tremendous blow to the prestige of Oracle as a company.

Now, most Oracle revenue comes from sevices and various accounting programs-but there are also starting to emerge various Open Source Accounting packages(i.e. SQL Ledger) that might in time start to hit Oracle more directly in the pocketbook.

Re:Is Oracle doomed?

[ckaminski]

Why do you think Microsoft is trying so hard to get .Net-ified? They want you to think of them like you think of Staples and OfficeMax, not as a software vendor, but as the underpinning of your corporation.

Oracle has the added deficit of not having an OS-level lock-in in the back office like Microsoft does.

"rollbacks" are an advanced feature?

[markv242]

I'm sorry, but that's like saying catching exceptions in your JSP code is an advanced feature, and you don't really need to use it.

Commit/rollback is an essential component to any decent data management system. Until you are absolutely sure that your data is correct -- that is to say, until you have done all of your transactions on the page successfully -- you should never actually write data to your database. Without using commit/rollback, you are stuck with the haphazard method of trying to manage potentially disastrous records of data showing up in your db.

Your other quote: "And MySQL will blow the doors off of Oracle and other databases in terms of raw speed" is similarly incorrect: MySQL may be faster when you are dealing with a small amount of connections, but as soon as your application starts getting any amount of concurrent users, MySQL is famous for falling down rather rapidly as it strains to write its data to disk.

MySQL cannot scale reliably, period. Having two database systems act as a pool, under MySQL, is a crapshoot at best. Unless you like designing single points of failure into your web applications, stay away from MySQL.

Simply put, if you expect your web application to get any amount of decent traffic (say 100,000 pageviews+ per day), then MySQL is simply not an option. Oracle is simply the standard upon which others can only attempt to compare themselves to. MySQL may be fine for the low-end "check out my k00l dynamic site!!11!!" crowd, but for professional web applications, MySQL has a long, long ways to go.

Re:"rollbacks" are an advanced feature?

[bad-badtz-maru]

Quizilla uses mysql and serves 300,000 pageviews per day.

maru

Re:"rollbacks" are an advanced feature?

[Alioth]

Simply put, if you expect your web application to get any amount of decent traffic (say 100,000 pageviews+ per day), then MySQL is simply not an option

How many pageviews a day does Slashdot get? I wager many more than 100,000.

Slashdot uses MySQL.

Re:"rollbacks" are an advanced feature?

[Graelin]

Simply put, if you expect your web application to get any amount of decent traffic (say 100,000 pageviews+ per day), then MySQL is simply not an option.

FYI. 1.5M per day - we run MySQL. It has and continues to run like a champ every single day for the last 2+ years.

Of course, we've thrown some pretty high-end hardware at it to keep it running this long.

MySQL cannot scale reliably, period. Having two database systems act as a pool, under MySQL, is a crapshoot at best. Unless you like designing single points of failure into your web applications, stay away from MySQL.

Nail on the head here. InnoDB (the real seller for MySQL, since it gives them ACID compliance) *really* sucks under load. It starts chewing itself apart. Funny to watch, not funny to clean up - since you really can't.

We've been using replication in MySQL for backup purposes. (The replication has always been reliable for us) We can take the slave down for snapshots. But that is *all* we use it for.

MySQL will last us just long enough to finish our PostgreSQL migration.

Re:"rollbacks" are an advanced feature?

[sloth jr]

Simply put, if you expect your web application to get any amount of decent traffic (say 100,000 pageviews+ per day), then MySQL is simply not an option. Oracle is simply the standard upon which others can only attempt to compare themselves to. MySQL may be fine for the low-end "check out my k00l dynamic site!!11!!" crowd, but for professional web applications, MySQL has a long, long ways to go.

My sites aggregate 62 million page turns a month, running MySQL - my largest customer often turns 300,000 to 400,000 page turns in a day. MySQL does fine.

Which is not to say your message doesn't have some validity. What we've found in general is that up to fairly high load, MySQL performance absolutely DESTROYS Oracle running the same application. After that point, of course, MySQL drops off a cliff, while Oracle just slows down. Without an exception, customers we've migrated to MySQL to Oracle are MUCH happier with their performance (again, because most of our customers aren't at that critical performance apex of MySQL). It's like this: 95% of everything our customers want to do can be handled with MySQL's functionality, at a mere pittance of Oracle's TCO. How much is that 5% worth, do you think? Now, this is a specific workload I'm talking about here - MySQL's peg definitely won't fit every hole our there, but it does fit some very important ones for us.

It's funny the reactions here - the folks who insist that MySQL isn't beginning to eat Oracle's lunch. It's funny because MySQL IS beginning to eat Oracle's lunch, for SOME APPLICATION LOADS. It's been pretty common for companies to just buy Oracle, because that's What You Do, especially when you've got high-power DBAs calling the shots. Many classes of business applications don't require Oracle's power, and savvy CIOs ARE waking up and specifying cheaper, more appropriate RDBMS for their applications.

Re:"rollbacks" are an advanced feature?

[scrytch]

Simply put, if you expect your web application to get any amount of decent traffic (say 100,000 pageviews+ per day), then MySQL is simply not an option

Sure it is, if you're running a blog like slashdot, which is 99.999% reads of the same few dozen datasets over and over.

Get more than a hundred or so simultaneous writes to a MySQL database and watch it just fall over dead, or queue them up like breadlines in SOVIET RUSSIA (sorry, that just slipped out). Heck, I don't even think it even has any deadlock detection.

Re:"rollbacks" are an advanced feature?

[mabu]

I'm sorry. Call me old school, but "rollback" is a feature for people who don't idiot proof their input routines properly.

Missinfg Features

[juergen]

As long as MySQL doesn't conform to all of ACID, it won't be used by serious players. So all those who use Oracle etc. and need a real RDBMs won't even try to switch. There was a lengthy discussion (or should I say ranting) over this in user comments in the online MySQL manual, but it looks like they removed that. Here's the best link I could find: Manual/ACID.

All those who can live with less, well, IMHO having these features still makes development of sound applications so much easier it pays off having it. PostgreSQL has most of Oracles features, conforms fully to ACID, costs the same or less as MySQL (nothing, compared to MySQL which is virtually useless free without the commercial table handlers), and there are some companies supporting it too.

In my experience an application which does correct error checking and handles faults etc. is not faster in MySQL than in most other DBs, just harder to write. And there are alternatives to PostgreSQL, if you don't like it.

Jürgen Strobel

Re:Missinfg Features

[Anonymous Coward]

Yea, but lets remember that not everyone who needs *some* form of DBMS solution needs a full blown RBDMS like Oracle. The whole point of this article is that MySQL does a dandy job doing what it does best: being just-slightly-better than a flat file data structure.

No, it doesn't do transactions, doesn't do stored procedures, doesn't even have referential constraints for $deity's sake. But that doesn't matter when you're programming a cheap web app where 5-10 extra lines of code in Perl of PHP can make up for it. Is it the best way to do it, strictly speaking? Probably not, but it works! And it works for really really cheap.

Who's to say that's such a bad thing?

It depends on what you want to do

[vinyl1]

I work at a heavy-duty Oracle shop. I would say that Oracle has gone way beyond being just a database vendor, in that they provide a complete--but proprietary--environment. Since I haven't needed to use many of their features in the past, I had never realized how complex their software is.

I can't believe MySQL doesn't even have subselects yet. I've been living on subselects and 'connect by' for years. I never did like PL/SQL all that much, but it does allow you to run complex programs over the network without creating high traffic. And SQLNet does make things a lot easier. The whole thing is kind of like a database flavor of Unix, with its own world of commands, scripting tools and permissions.

As for Oracle's 'fancy' products, like Express, Forms, the OID, Portal, and Workflow, they are serious attempts to extend the database principles into a generalized suite of enterprise-level business tools. They are a little too cutting-edge for my taste, but you won't find anything like this in a non-proprietary product.

Must be a slow news day....

[miketang16]

We're getting articles from CNN now??

Why not mySQL?

[Anonymous Coward]

Philip Greenspun wrote a short and excellent article on ACID compliance. The article is 3 years old, yet mySQL still has problems as they developers don't seem to believe that ACID is important. Open ACS on "Why Not mySQL"

mySQL is, unfortunately, a SQL interface to a bunch of files based on various index sequential access methods. It gets its speed by ignoring transactions, triggers, stored procedures and other things that, when your company is successful, will need in its database. mySQL's replication is also not guaranteed and when its spotty, it doesn't tell you.

The open source DB community is a powerful force with a lot of potential and a lot of success. That success is in markets where transactions are low and/or not critical to the customer.

mySQL and others need to ensure that they have these features:

• stored procedures (implementation outside of the A in ACID aren't complete - perl, java, python, etc)
• Referential integrity, foreign keys, transactions
• hot backups where you don't have to take the database down to get a backup with guaranteed integrity.
• reliable replication (argue away, only shareplex, NT SQL server & Sybase have it today)
• sub selects
• temp tables
• function based indicies
• automatic partitioning
• rollback (true rollback w/transactions)
• triggers
• block and row level locking. A select on a 50 million row table shouldn't lock the table.
• joins that do not lock tables due to full table scans

There are a lot of good reasons for using mySQL as a platform to begin the development of a project. For personal use, it's hard to beat! If you are a professional in a company that needs to support real clients with real data with real guarantees, spend your money on a real database.

Where do you want to spend your R&D money? On your product or on the database that does most of the things you need, but not all of the things you need. Don't you want to spend your time building the product that pays your salary and makes your customers happy? Why spend time on the database, just buy something that works.

One more thing, a not unreasonable architecture for a database driven application is:

• UI layer
• Business rule/application layer
• Application Programming Interface
• Stored procedures (potentially hundreds)
• Database

Good luck.

Re:Why not mySQL?

[Hrunting]

The article is 3 years old, yet mySQL still has problems as they developers don't seem to believe that ACID is important. Open ACS on "Why Not mySQL" [openacs.org]

Well, the ACID qualifications have been satisfied by MySQL through the InnoDB handler for over two years. Other issues, such as advanced SQL features, are still in progress.

But to the original point, I think it's a good thing that MySQL doesn't think that ACID is important. They have different priorities. Their priorities do not lie in making a database that conforms to someone (anyone) else's theories of the perfect database. They implement features that are requested by their customers (pay a minimal licensing fee and you suddenly get a much greater voice in the direction MySQL is taking) and their overriding goal is to implement those features as efficiently as possible, so that speed and performance aren't sacrificed for feature bloat. They tend to implement their features along the line of standards, but they also through in numerous "extensions" to that standard that make way for faster processing and quicker development time. That sounds to me like an excellent way to run a software project.

With that said, they have a roadmap for implementation of various features, including subselects, triggers, etc. and they've already explained why one is getting implemented before the other. If you think a really, really important feature is missing, pay them some money and then say, "Hey, this is a really important feature to us and we just gave you $XXXXX."

Re:Why not mySQL?

[tsmoke]

Philip Greenspun didn't write this. Ben Adida did, and it's a big difference.

Philip started arsDigita and the arsDigita Community System (ACS). Ben started the Open arsDigita Community System community (OpenACS).

Ben wrote the article to answer people why the community worked to promote PostgreSQL and not MySQL.

talli

Re:Why not mySQL?

[fferreres]

# block and row level locking. A select on a 50 million row table shouldn't lock the table.

As far as I know, I SELECT doesn't block the table. It blocks it when you do an INSERT or UPDATE.

No. It isn't.

[philovivero]

MySQL is not a threat to the bigwigs, because they compete in different realms. MySQL is a threat to filesystem-storage and BerkeleyDB.

PostgreSQL is a threat to the bigwigs, however.

This is not to say it won't change. MySQL apparently is trying to implement features that would make it compete with real relational databases, but last I heard, views weren't on the list, so I'm not holding my breath.

Other OSS projects that may be a big threat include SAP DB (used to be Adabas D) and... uh... right. There you go. Reply if you're a real DBA and think there's another competitor in the space of true relational RDBMSs. Hint: If you think MySQL could be on the list, you're not thinking of industrial strength databases.

Re:No. It isn't.

[WasterDave]

I'm not a real DBA, but know one - and he messes around with Sybase, of which there is now a free version. Probably only free as in beer, but hey.

Me, myself, bollocks to it all - standardise on postgres and be happy. It's gradually becoming less of a bitch to administer and has never had any problems on the stability front.

Dave

Re:No. It isn't.

[philovivero]

WasterDave, eh? Heh. Okay. I know you!

Sybase definitely, if you use 11.0.3.3. But keep in mind that 11.9.2 and 12.0 aren't free (as in beer or speech) except for development. For production, you need to purchase it.

Sybase 11.0.3.3, which many would say is production level (although lacking row-level locking...) is free (as in beer and only on Linux) for all usage, but you don't get source code, and I'm fairly certain it's not too well-maintained anymore.

And since you're the guy who mostly turned me on to PostgreSQL in the production arena, I prob'ly'll just defer to you on that point. No matter how you slice it, you've got the source code and you can get commercial support. Nice.

Re:No. It isn't.

[philovivero]

re: firebird.sf.net.

Thanks for the link. I knew I'd heard of another RDBMS, but couldn't remember it.

I don't know anyone who uses Firebird, and I've never used it myself. History says it came from Interbase which, again, I've never used nor known anyone to use. So it's a complete unknown to me. Not saying this is a statistically-relevent fact, I'm sure a lot of people have used it, but I started out my career in a Sybase shop and moved to an Oracle career path, currently hovering somewhere between Oracle and PostgreSQL.

If I were recommending a database to a customer, the known quantity of PostgreSQL would admittedly make me extremely wary of recommending anything else. For me, until proven otherwise, I'll probably recommend:

0. Ultra-simple SQL needs, mostly read-only? MySQL
1. Simple SQL needs? (Pure datastore but transactional nature): PostgreSQL
2. Simple datastore w/ Stored procedures? PostgreSQL
3. Complex RDBMS needs? Perhaps Oracle if you can stand the pain, but Sybase will save you some cash, and PostgreSQL might still be up to task.
4. Ultra-Complex RDBMS needs (replication and database clusters)? Oracle.

This is all OLTP. When you start talking about OLAP, you gotta start thinking all over again.

Re:No. It isn't.

[WasterDave]

Real DBA. See, told you.

Dave :)

Re:PostgreSQL has every feature but Replication.

[UnderAttack]

Not sure why it hasn't gotten around to PostgreSQL yet that MySQL does support transactions.

I see it as one of the main advantages of MySQL over PostgreSQL is that you are able to turn off transactions if you don't need them.

The main difference between MySQL and PostgreSQL is more 'philosophical'. MySQL does not attempt to hunt Oracle based on features. Instead, the main objective for MySQL is speed. PostgreSQL on the other hand attempts to duplicate as many Oracle features as possible.

BTW: MySQL does support replication well, even in its non-commercial version.

Threat is in the mind...

[MoThugz]

When viewing from a pro-Open Source perspective, yes you can argue that MySQL is maturing well enough and hopefully can challenge the bigwigs in a few version revision. This argument is not only targetted at MySQL, but other open source DBs such as PostgreSQL, SAP DB, etc.

This is because in out typical Slashdot environment, people are supportive of the Open Source movement. But companies (suprise, suprise!), do not follow our line of thought.

We see it as free, they see it as if they are giving it away for free, then there must be a problem with it. Or more precisely I'm not gonna use it because there won't be anyone who's available 24-7 so that I can ask support about something which is already in the docs but I'm too lazy to go through it.

Some may laugh, but if you are working in a technical field... you'll start to wonder "Why the hell did I spend 3 days working on this documentation when the lusers won't be reading them anyway!".

And guess what, the big corporations know this! They'll say that with their product, all you need to worry about is what to use it for. Installation, servicing, patching will be covered by the company. Furthermore, they'll bundle application design apps, etc., to make your life easier. With an open source DB, you'll have to code one yourself, or use another that's already available.

We have to think as lusers instead of coders, to see why this bigwigs are getting paid a whole lot more for releasing something which is already available (to a certain extent) for free.

mSQL

[felipeal]

Of course, there's no mention of PostgreSQL or mSQL, but I guess that's typical.

Not being a troll here, but does people still use mSQL anyway? I know it was very popular a couple of years ago, but nowadays I guess it lost its place to mysql...

Surprised....

[PrimeNumber]

I am frankly surprised that MS SQL Server was ranked along oracle and DB2 as a 'high end' DB. Anyone who has had to work with it usually disagrees!

Personally I have seen SQL server most on small/medium size business environments. Any large 'enterprise' sized business deserves what they get if they are dumb enough to rely MS SQL server. Look what happened to Bank Of Americas ATMS when the last MS virus du jour made its rounds.

I think MySQL is the best bet to reduce Microsofts share of the DB market. Oracle is better, but small business isn't willing to fork out that kind of cash, especially in this economy. MySQL is especially perfect for the small business web site, and with Microsoft irrationally increasing subscribtion fees and forcing upgrades, a good percentage of their customers will be running into the open arms of MySQL/Postgres.

Re:Surprised....

[Seahawk]

I have no worked with oracle and db2, so i cant compare it to that - but i use mssql professionally and mysql personally, and comparing the 2 makes me believe you have no idea of what you are saying!

And how that got modded to 4 is beyond me!

And a little note on Bank of America - people that doesnt patch their servers for 6 months will get into trouble - no matter what os they use!

Re:PostgreSQL has every feature but Replication.

[tzanger]

I got around the replication issue (mirror only) by putting a insert/update/delete trigger on the table I wished to replicate. The function simply logged the action into an "audit" table.

That audit table had a delete trigger on it which wrote the deleted tuple to a third table. A cron job ran every five minutes which executed 'DELETE * FROM table_audit;', pg_dump to a textfile and finally 'TRUNCATE TABLE table_audit_copy;'. The resultant textfile was compressed and sent over to the other database, where some postprocessing turned the audit logs into proper queries again and executed them on the mirror DB.

This "double-buffering" is necessary in order to eliminate consistency problems in the database audit log. (e.g. if we just dumped the audit table we wouldn't be able to easily tell which entires were successfully dumped on the most recent pull since new data is coming in all the time, and locking the table while dumping it causes access delays on the table I wanted to replicate.) Using triggers and letting the DB handle it itself is a very clean and efficient way of handling this kind of data consistency problem.

And yeah, I know this isn't the best for performance and doesn't suit everyone, but it does seem to work well. I could have written a function to just log the tuple that changed in the main table to disk directly, but just like perl, TMTOWTDI. :-)

As it turns out, I am not even using this anymore, since it was easier to modify the client that was doing the database work and have it just log the transactions to disk instead. (They were insert/update only.)

Something is also nagging at me that there were some transaction problems with the default transaction isolation level, but I have since forgotten the issue.

I think

[djupedal]

...I'd like to see the stats where SQL (big boys) is oversold. In other words, where is it being used as clear 'overkill'.

The odd thing about MySQL

[harlows_monkeys]

The odd thing about MySQL is I've never seen anyone give a good reason for reason using it in the first place.

As far as I have been able to tell, every reason one might give for using MySQL also applies to PostgreSQL.

On the other hand, there are several things PostgreSQL can do that MySQL cannot.

So...starting from not having any database, and deciding which to use on that first project, I don't see why anyone picks MySQL over PostgreSQL. They seem to be about equally difficult to set up and learn, so why not go with the one that will be able to handle more? Even if you don't need that more now, why not, since it is no extra effort, be prepared if a project comes along that does need those extra things?

Re:The odd thing about MySQL

[jenkin sear]

In my experience, MySQL has been faster at doing select-type operations. I use it for a web-based CMS, where transactions are unimportant, and the majority of database work is grabbing stuff from the DB and displaying it.

I use postgres when I'm concerned about data integrity, and speeding up writes to the database- if I'm doing a sufficiently complicated write to the DB, postgres' stored procedures make it a much better idea. I've used it for embedded-type monitoring and data collection applications.

So there are situations where one or the other is better- like sometimes perl is better, sometimes C is better. Different tools for different jobs.

Oracle is preferable to me & my company.

[BoomerSooner]

However we use MySQL, MS SQL Server and Oracle (different solutions/architectures). I don't see the pricing problem with either MS or Oracle. You can get the standard editions for around $1500.

Sometimes paying a little can save lots more in the long run. I personally hate Transact SQL but it's probably because I started on PL/SQL first.

However, that being said I just don't understand some of the stupid ass implementations in TSQL.

MySQL is great because it's small, cheap(free), and very reliable (in my usage). Plus it's got that great JDBC driver that some guy wrote (name?) and the MySQL people hired him (way to go buddy!).

Oracle just seems to have the whole ball of wax. I've never felt I'm trying to program (stored procedures) around a piss poor implementation (like TSQL). So for my (and my companies) money I still say Oracle is the way to go. Plus, I have yet to see the reason to migrate from 8i to 9i.

I'm seriously considering implementing Oracle Financials as well. Show me the MS or MySQL product that can compare to that!

Re:Oracle is preferable to me & my company.

[fitten]

Heh, I have no problems with either PL/SQL or T-SQL. There's no magic behind either one.

My main issue is featureset. For things that we do, MySQL just won't cut it. It gets closer but as slow as some of the features are put in, I might be an old man before it can be usable for what we do.

Re:Oracle is preferable to me & my company.

[Mad Browser]

Watch out for Oracle Financials. In our experience, they promised us the world but were never actually able to demo anything beyond the most basic accounting features...

Re:Oracle is preferable to me & my company.

[Malc]

Uhhh, what's wrong with T-SQL? You diss it, but don't point out any of its problems.

Re:Oracle is preferable to me & my company.

[Captain Large Face]

Plus it's got that great JDBC driver that some guy wrote (name?)

Mark Matthews?

Linux Registry?

[SHEENmaster]

As I remember a certain evil corporation attempted that. They result was that apps would store everything from logs to movies in the registry which resulted in major ram depletion and instability.

Re:Linux Registry?

[RevAaron]

See the other poster's comment about this being like MS's registry. It is not. Any problems with the Windows registry is an issue with implementation, not the concept in general. There is an equivalent to the registry available for Linux, used by some GNOME apps. Not sure how many actually use it (%) or what it is called at the moment.

Nonetheless, the kind of system-wide database is not the same thing as the MS registry, although the MS registry is a good idea, although (very?) poorly implemented. I'd much prefer it to having flat files that are in scores of different formats. There is no reason you could not edit this registry from the command line using a small utility just like you could edit flat files.

Replaceing MS Access

[rossz]

I'm investigating what is required to replace MS Access. Yes, some company made the mistake of basing their important information on that piece of crap.

I plan to use MySQL (probably 4.x) for the back-end. For the front-end, I might use OpenOffice. It seems to handle forms ok. Hopefully, I can figure out how to produce a decent report system in OpenOffice (you'd think that would already be built in).

I've tried to find an open source form/report system, but nothing I could find seemed adequate for the job. Commercial packages were always way too expensive (thousands of dollars).

Any suggestions?

Re:Replaceing MS Access

[occamboy]

I know that I'll get yelled at by this crowd, but, without knowing more specifics on your application, I'd recommend PostgreSQL DB, a Visual Basic front end for forms, and perhaps Crystal Reports for reporting.

This will give you a nice blend of tremendous power and ease of use.

Re:Replaceing MS Access

[PrimeNumber]

Since its access/users/mysql you are talking about I might be able to help.

I am going to assume the clients/users are still using windows. MyODBC can be used to connect to your database. (BTW ODBC = Open DataBase Connectivity.) Then you can use MySQLfront to help create tables, etc in a graphical UI on a MS windows box. This means your office 'power users' wont be too mad they wont have their access wizards. ;)

Now that you can connect to a MySQL DB, and users have gui tools to create tables and select statements, then you can use something like occamboy suggested (Crystal rpts) if you are stuck using VB. OR you might try using a browser as a reporting tool. This works good if cash is limited.

Example: Using php/apache to build the report html, make a table or DIV for each page, setting a strict table/div width, then at the end of pages insert style sheet page breaks. The user then can print directly from the browser. Yes it *will work* with IE if you are still stuck with it. BTW I would recommend setting the permissions different on the reporting tables, so users can't cheese 'em.

Hope this helps.

Re:Replaceing MS Access

[rossz]

I've already been fooling around with myodbc. There seems to be a bug in it wherein you must give superuser access to the connection if you want to make changes to the database (not good). I'm still fooling with it so I haven't verified it isn't a mistake on my part.

I just downloaded MySql-Front and will play around with it. Thanks for the info.

I considered a web interface output for the reports, but rejected the idea because printing is far more important than viewing with a monitor and getting exacting printouts via a browser is a pain in the ass.

Worse case, I can always use Crystal Reports. But I would feel so dirty if I used Visual Basic to tie it all together!

I'm looking into Report Manager from Borland. It might do the job, though so far I'm not very impressed with the interface (a bit of a pain to use).

Re:Replaceing MS Access

[rossz]

Not a chance. I refuse to give Microsoft any more of my money. Besides, I'm trying to use this project as a way to show that open source is fully capable of doing the job.

Someone would have a popular package if they wrote a nice open source replacement for MS Access's forms and reports. Despite being a rather shitty sql program, MS Access has a very nice implementation of forms and reports.

Re:Replaceing MS Access

[tqft]

Trail of tears: MySQL, ODBC & OpenOffice 1.0 http://www.osforge.com/news/00941.html "... McCreesh writes about OpenOffice.org 1.0's "best kept secret" -- that secret being the fact that hidden away inside, completely unknown to most OpenOffice users, is a user-friendly front end for databases that is "a Microsoft Access (and more) equivalent." That may be so, but there is a very good reason why it's a secret: it's too damn hard getting OpenOffice and ODBC wired up correctly.." This article explains how to get it working

Re:Replaceing MS Access

[rossz]

My, what a powerful and eloquent retort.

Listen, Fucktard (to use your own words), MSDE is not free. It is included with other products which you must purchase.

which one?

[smilo-bob]

Has anyone successfully done this with any of the open source DBMS's ?

* Created a database with 40G of data capacity
AND
* Kept is online 24 x 365
AND
* Created backups that were 100% guaranteed to restore the database in the exact state of when the backup was created
AND
* allowed 20 or more concurrent DML conections to the same TABLE (but different rows) without blocking
AND
* supported ROLLBACK and COMMIT
AND
* supported subselects
AND
* supported a stored (in the DB ) procedural language
AND
* received reasonably good performant

If you answered yes to ALL the above, what was the name , platform, and release of the software?

Thank you

Re:which one?

[AWrinkler]

Firebird (was Borland Interbase)
http://firebirdsql.org
it's had all these abilities for years.
I've had 76GB single-file databases on my FreeBSD machine since last year.
Faster than Postgres on everything but deletes, but it cleans up after itself when postgres just marks pages for deletion during the next sweep.
Same speed as MySQL on insert/update/delete.
Slightly slower on selects, but that's understandable.

After all, it does have:
Stored Procedures and Triggers using the same PL.
Views, cursors, custom datatypes.
Multi-terabyte file handling capacity
Transactional Engine, with full Commit/Rollback.
Full referential integrity

I've had it doing ~300 transactions/sec on my Celery450, so it rocks along nicely.

I've used Mysql, PostgreSQL, SQL Server.
So far, Firebird outstrips them when you weigh all the features and performance.

Sure, you can power huge sites with MySQL, in the same way you can pull a 6-berth caravan with a Daewoo.

Firebird is a 4MB install. Why choose anything else?

BTW. It works a dream with PHP, Perl(DBI), C.
Probably more, but they're the ones I've used personally.

Re:which one?

[Anonymous Coward]

We had a DB with about 65GB of data. It is being used 24x7, and we do backups every 12 hours. We have around 200 users all the time over about 85 tables. We use transactions alot, but I can't be certain on the number of ROLLBACK's that are issued. Plus we use the internal postgresql stored programming language.

It is an internal application for an insurance company. We run the server on HPUX, currently on version 7.3.1 of postgresql.

It isn't too hard to find people using postgres for serious work. Just hang out on the mailing lists for a week and you will be able to write a short book on them

Re:which one?

[javabandit]

Actually, multi-valued/post-relational databases have been doing this for many years now. Such database vendors are Pick Systems and IBM uniVerse (formerally Informix/VMark).

These databases are extremely fast, loosely typed, support many different table types, support SQL, support multi-valued (columns within columns) versions of SQL, fully ACID, have APIs written for every major language out there. All of these databases usually support two procedural languages underneath which are native to the database. One is called Access/TCL (Terminal Control Language). The other language is ordinarily a flavor of BASIC. The former language is extremely terse, much like RPG. While the latter is much more like writing a GW-BASIC program.

These databases have been around for decades now. They are running in huge environments, processing terabytes of information, have been tested and re-tested, and have been stable for decades.

Check them out if you are interested.

Re:which one?

[javabandit]

You are calling "multi-valued" "post-relational"??

That's not what I call it. That is what the vendors of these databases call it. Because these databases are RDBMSs that support multi-valued relations as well as 1NF relationships.

Currently, the only database management system (although the current implementations are 100% it), based on a foundation of LOGIC is the RELATIONAL Database Management System.

Again, you should do some reading yourself. Multi-valued databases also support the same features as 1NF RDBMSes. If you want 1NF, by all means, you have it. But if you want multi-valued, you have it there, too.

Multi-valued is just another word for "eh well, put anything in it - I don't care".

Again, you're wrong. Multi-valued databases provide an easy-to-use mechanism to denormalize data when necessary. And though all of us would love to use 1NF 100% of the time, it simply is not reasonable if you have huge amounts of data that have to be repeatedly summarized, or sliced-and-diced.

Learn before you speak, at least.

Re:which one?

[ron_ivi]

Sure... the guys running the .info domain (and now .org) on postgresql have been doing this with much larger databases than 40G.
http://cw.idg.com.au/idg2.nsf/All/2ADD84E6EBCEADE9 CA256CB30075FA01!OpenDocument

Typically ...

[DogIsMyCoprocessor]

a human estimate ("I estimate we command 20% of the world wide installed base of databases.") is way off. A lot of Windows applications run on top of Access or SQLServer, which probably puts the real figure at something more like 2%.

You're overlooking many other database engines

[Jeddawg]

This discussion seems to be omitting an entire market segment of database engines. For instance, Advantage is a powerful database server that's priced at less than half the cost of M$ SQL Server, and Oracle. It's not a "free" product, like MySQL. In addition, Advantage isn't limited to SQL, which is nothing more than a limited reporting language, meaning, it's much more flexible to utilize Advantage than either of the products you seem to be comparing. And, No, I don't see MySQL as a threat at all to products such as Advantage.

In addition, MySQL isn't really all it's cracked up to be. Features such as page-level locking, (used by MySQL) and locking escalation (used by M$ SQL) will degrade performance in a multi-user application. So, while MySQL is great for a web-server, developing/deploying an application that uses MySQL can cause undesired performance degradation when multiple users are simultaneously accessing the data.

SQLite

[ddkilzer]

SQLite claims to be twice as fast as both MySQL and PostgreSQL, and is more SQL92-compliant and ACID-compliant than MySQL.

Does that mean everyone should drop MySQL and PostgreSQL for SQLite? No. It means you have to evaluate your situation and choose the best tool for the job.

Personally, I've have very good luck using PostgreSQL, and probably won't ever consider using MySQL until it is truly ACID-compliant.

My Story

[SloWave]

A couple years ago I started a large hardware conversion project for a major telco. One of the requirements was a fairly large database to support real time call processing. I had already told the customer that I would only do the job if it was on a non-Microsft platform so I didn't need to worry about them wanting SQL Server. However, since they were a large telco I assumed that they wanted a well know commercial product so I proposed either Oracle or Informix - their preference. Their director of DP said something like "it's too bad we can't use MySQL" since they were using it for some smaller applications, unknown to me. My next comment was "do you want to use MySQL?". The answer was "yes, provided it could do the job". I said "I will make it do the job". Now it's been about two years and MySQL has almost faded into the background. It just runs, unlike my experiences with Oracle and Informix where you have to constantly administer them. That's my personal experience, your mileage may vary depending on your skill and attitude.

MySQL DOES have sub-selects

[laptop006]

I've seen it demonstrated, and tried it myself, although as the MySQL guy pointed out a log of sub-selects can be done with joins.

The next version of mysql will also have views.

Re:MySQL DOES have sub-selects

[vinyl1]

No views? This is quite a serious limitation. I suppose you can't define network objects, either.

This is where Oracle shines. Using views, network objects, and synonyms you can do amazing things.

Of course, you can also make everything so complex that no one will ever figure it out...

Re:MySQL DOES have sub-selects

[micromoog]

...as the MySQL guy pointed out a log of sub-selects can be done with joins.

This is a pathetic excuse for a MAJOR lack of functionality. If a query can be rewritten to avoid the sub-select, why didn't the MySQL optimizer just do so?

Until a database can support transactions, subselects (yes on UPDATE/INSERTs too) and views (one of the most fundamental relational features) it has no business pitching itself as a real relational DBMS. MySQL is a glorified file system, and works well for people who need a SQL-like interface to a fast file system.

Re:MySQL DOES have sub-selects

[zenyu]

MySQL is a glorified file system, and works well for people who need a SQL-like interface to a fast file system.

Yup, nothing wrong with that. It's like BerkleyDB except with the same SQL interface so you can test your queries without compiling. Pretty neat. It's also fast, which is great. You need a good DB admin to get performance out of Oracle, and mostly you get that by avoiding stored procedures and views... I worked on a n-tier system where one version major version release was simply eliminating 90% of the stored procedures and the view based "Security" system. These were all moved to a Java program on a server sitting next to the DB server. They had a fast connection so all the extra data that had to be sent because we didn't use views anymore never made it out of the rack. A huge performance boost (I don't think our DB admin was the best...). I don't think that company will ever move away from SQLServer/Oracle unless a customer requests it, just cuz you stick with what you know works. But it could be done, in n-tier the DB doesn't need much smarts. (Well ACID, but they say that's coming to MySQL, and is in PostgreSQL...)

Still there are a LOT of places where flat files are still used and BerkleyDB as well, just replacing BerkleyDB with a 'SQLish' DB is a major win for 'real databases.'

Re:MySQL DOES have sub-selects

[laptop006]

As I said MySQL 4.1 does subselects that can be used almost everywhere, take a look on the mysql website for details. And the views in the next mysql will work even for updates and insterts.

Re:If MySQL was just a bit more user-friendly...

[chriskl]

Ummm...Oracle does.

Hardforum

[BHearsum]

I'm surprised this hasn't been mentioned yet, but about a year and half ago a big gaming/computer forum was running MySQL with vBulletin as the frontend. Horrible permorance. They had top-end Xeon servers dedicated completely to MySQL, and they still had about an hour of downtime everyday. I'm not sure if this true all the time, but in this case, MySQL could not handle heavy loads. They had to end up taking out there 'General Mayhem' forum, where more than half the traffic was.

Not really

[The Man]

mySQL is appropriate for upwards of half of all web applications, and could easily own that half of the market. However, that doesn't mean it constitutes a serious threat to the large proprietary database vendors, because that half of the applications are mostly using one or more of:

• Microsoft Access
• Flat files
• XML files
• Static content
• Client-side scripting
• A large-scale database being drastically underutilized

to perform their various functions. In most cases, those functions would be faster, easier to implement, and simpler to manage using mySQL.

For applications with these types of functions, which do not include complex queries, large transaction volumes, rigorous reliability including transaction log backups, recovery, replay, and replication, mySQL represents a major force. Unfortunately for mySQL and those who would have it take over the world, there's not much money available for those applications. Therefore, expect to see mySQL's installed base continue to increase while its revenue-based market share remains small.

For applications which do require features and levels of reliability and capability not offered by mySQL, postgres is the only serious freely-available contender. Even so, postgres is also somewhat less capable than Oracle or DB/2 and will be confined to the middle tier of applications - those which require better reliability and scalability than mySQL can provide but for which funding is scarce. Postgres probably does represent a serious threat to Microsoft's SQL Server, if only because Postgres is platform-independent and supports platforms which can scale beyond anything Windows can run on. Both are otherwise middle-tier products which are not and will never be taken seriously by the largest and most demanding database users.

Who are those users? Banks, government agencies, stock exchanges, payroll and records processing firms, insurance companies, large multi-site call centers, and other huge-scale enterprises. The top proprietary databases offer capabilities that do not yet exist in the Free Software world. For these users, who are less than 1% of all customers but which represent maybe 80% the revenue in the market, there is no substitute. These customers will stay with their existing solutions - Oracle, IBM, Sybase - until the systems running them give out. Then they'll call that company's professional services department and offer them a few million more to upgrade the system. That's the way it works. The system has to be attacked one customer at a time, an expensive and time-consuming process consisting of many lunches, legal bribes, and unrealistic promises.

I think the answer to whether mySQL is a significant threat to dominate the market economically is pretty obvious. Even if mySQL moves up to the middle tier to compete with Postgres and MSSQL and is installed in every application for which it is suitable, the product would still command less than 10% of the revenue in the market.

What a silly question.

Re:Not really

[tobes]

All it takes is for one or two of the big guys to pick up Postgres and bring it up to the level of Oracle. All of the cash flying around for those Oracle licenses could sure fund a lot of oss db development.

Re:Not really

[The Man]

Yes, but they don't think that way. Remember, the Oracle guys take them to lunch every day, and let them sit in their skyboxes at football games. Where's the incentive to develop alternatives? Worse still, the people the big guys use for development are usually not what we'd think of as competent; typically they are consultants from other big firms like PWC or Accenture. Most of them, given the source for postgres, would stare blankly into space for six weeks, add on a few crufty modules, and vanish leaving no documentation. Of course, the big guys *could* hire some people with existing postgres experience or even a couple guys from Oracle or IBM, for a lot less than they would spend on the next Oracle project, but they won't. Why? Two reasons: first, the guys from Accenture give them tickets to baseball games and golf tournaments and a lot of free lunches; second, Oracle, whatever problems it's given them, has been working more or less well, and sticking your neck out for change is a sure way to get your head chopped even if the new project causes fewer problems than Oracle would have.

The safe, easy, personally satisfying thing to do is exactly what's always been done - hire a large well-known company, maybe two or three, to install massively complicated, undocumented, hideously expensive proprietary software. Then hire them to maintain it. And enjoy all the free stuff they give you while you cover your ass and project the image of a real go-getter.

The mistake you're making is to think that anyone at the big guys who holds the purse strings actually cares about anyone but himself. The company is secondary at best and the prospect of helping the rest of the world by contributing back to Free Software projects means nothing. This attitude is a requirement for receiving the key to the executive washroom at a Fortune 1000 company.

Apples and Oranges

[SatanicPuppy]

I don't know why people insist on comparing MySQL and Oracle. Oracle is huge and bloated, but it runs pretty quick, and is chock full of the sort of features you need if quadruple redundancy and data integrity are a must. If I'm working for a company that can afford the licensing, I'm Oracle all the way...There is no commercial product that really compares.

On the other hand, if I'm dealing with a company that can't toss around the kind of money that you have to have for an Oracle DB, MySQL is my number one choice. I can slap the GUI of my choice on it, take care of data security with a hard backup and pocket a few grand of pure profit that I didn't have to spend on liscensing. You can argue Postgres, but I've never run into a case where I couldn't work around those features that haven't been implemented in MySQL yet.

The one thing I can't stand is when someone suggests: "I can't afford Oracle, so lets' go with a MSSQL database." That's like, I can't afford a space shuttle, and a ferarri isn't good enough for me, so I'm going to buy this million dollar llama instead because 1000 marketing agents can't be wrong, right?"

It has all the same feautres as Oracle, it's just that the features in Oracle WORK.

Just my .0363160 Bulgarian Leva worth

Yeah Right...

[praetorian_x]

And next you'll be telling me that Linux is a challenge to Solaris and Windows...

Er, wait a sec...

Cheers,
prat

MySQL is ACID

[Imperator]

MySQL does have transaction support and is fully ACID-compliant--iff you use the InnoDB table type. This also allows you to use foreign key constraints. However, it's not as fast as MyISAM and doesn't support certain features (e.g. fulltext indexing).

In my (informal) tests MySQL/InnoDB is less than half the speed of MySQL/MyISAM but still about 50% faster than PostgreSQL for simple and small tasks. That said, PostgreSQL has more features than MySQL and I still prefer it for most tasks.

Re:Ethical obligation?

[Tony-A]

Enlightened self-interest would be more accurate but "ethically obliged" looks more intelligible. If I do manage to run into something "interesting", I think I do have an ethical obligation to put out some effort to help get to the bottom of things. (After checking latest updates which will most likely have already solved the problems;) This is in contrast with commercial software where it's just not worth the hassle.
Too much like a pot-luck supper. While you can free-load, it's not comfortable feeling.

Get serious

[tmark]

no mention of PostgreSQL or mSQL,

Is this a joke ? Forget, for a moment, the conclusions we're supposed to draw from 1) the observation that MySQL may run a lot of websites (this is about as relevant as pointing out that Hyundais outsell Ferraris - doesn't mean that Hyundais are superior vehicles), and 2) a lot of commercial websites might run MySQL (also about as relevant as pointing out that companies buy more Ford Focuses for their fleets than Hummers - doesn't mean the Focus is a 'better' vehicle than the Hummer).

mSQL is about as far away from providing the feature set of MySQL as MySQL is from providng the feature set of an Oracle or PostgreSQL - which is to say, worlds away. Sure, MAYBE mSQL is the tool for a particular job - but then we have to ask, are flat files a threat to the mSQLs and MySQLs - hell, the Oracles- of the world ? After all, flat files are free and we know they're in wide use in lots of companies.

Have a look at SAP DB before talking about those..

[FeatherBoa]

features that MySQL (and other open source DB's) just don't have, and probably won't have for Years.

SAP DB is free, open source and GPL. It also has all the best big-guy features. Not many people seem to know about it - it certainly has small mind-share. But it is the real stuff - miles ahead of MySQL.

Other open-source dbms

[pspinler]

that have received little comment so far:

* Firebird (ne: Borland Interbase)
* SAP-DB (ne: Adabas-D)

Both are good, high quality, commercial or formally commercial products released under an open source license. (interbase public license and GPL respectively)

Further, SAP-DB has excellent commerical support available from SAP, the company, at or better than the same level of responsiveness as, say, Oracle support.

Both are fantastic, enterprise level full ACID RDBMS's with all the great management features a heavy duty shop could want:

* online backups,
* transaction logs,
* restore to point in time
* subselects, views, rules/triggers, procedures, etc.
* great storage management

Check 'em out.

-- Pat

Check TPC results

[melted]

Among the enterprise-level DBs, SQL Server beats the crap outta every other DB both in price/performance, price AND performance. So TPC says it's got the highest-performance and it's the cheapest DB available on the market. See http://www.tpc.org/ for more information.

How do they resell GPL code?

[sbwoodside]

I don't get it. MySQL releases under the GPL, but they also release under a proprietary license as well. They must require that patches be submitted with copyright assigned to MySQL AG ... right?

And Unicode

[GCP]

UTF-8 is the default encoding for XML. If you want to store default XML in a database, or create a database that can handle any text your customers submit without corrupting it, or that can handle all of your major markets without requiring a different DB (but with the same functionality) for each region...you'll have to use PostgreSQL, because though MySQL has been promising UTF-8 for years, they just don't seem to understand it well enough to ever deliver.

Re:And Unicode

[skillet-thief]

because though MySQL has been promising UTF-8 for years, they just don't seem to understand it well enough to ever deliver.

Unicode is announced in MySQL v. 4.1 (I think). Is that for real? I mean: will it really work?

I am currently using MySQL for a web-site. We need to introduce a bunch of new international alphabets (no! please no!). I'm not sure whether to wait for a UTF-8 version of MySQL or to just plain switch to Postgres.

Any thoughts?

The other shoe has dropped

[pvera]

I am a web developer and in the last few years I have written products that use SQL Server 6.5, 7, 2000 and Oracle 7.x and 8.x. MySQL has always been present whenever we discussed data back ends and was always dismissed as "not good enough." Usually because it did not cost a gazillion.

During the last 18 months or so I have run my personal sites with a MySQL back end. I have never had an outage or loss of data that can be traced back to the MySQL servers. I ran it first on Windows 2000, later on freeBSD4.5 and now on a freeBSD4.6 jail. It still works perfectly.

Back when we were still arguing (two jobs ago) about using MySQL, the DBAs usually claimed that you could not trust MySQL because of the lack of stored procedures and the fact that it could not pass an ACID test. Since then I never bothered learning the DB system itself beyond the minimum needed for SELECT/INSERT/UPDATE/DELETE operations, I did not try to verify this on my own. Years later and I am convinced that these DBAs probably read that in a magazine, and that none of them had even seen MySQL running.

After the dot-bomb nightmare nobody in his right mind should be proposing to their managers to spend obscene amounts of money in SQL Server and Oracle licenses just to do simple SELECT/INSERT/UPDATE/DELETE operations. Sure, stored procedures rock but it does not make any sense to spend that much money just for the 1% of your functionality that will be run by stored procedures!

Most of the people I know that use SQL Server don't even know how to write one, and in the last 5 years I have only written two web applications that have more than 1% of their sql operations as stored procedures. And for Oracle it is even worse!

Online Backups???

[germanbirdman]

As far as I know, both Postgres and MySql do not support online backups.

This makes it unusable for a lot of enterprises.

Also, many people here do not seem to be aware that MySQL with InnoDB tables can do commits/rollbacks and in the last C'T [German Computer magazine] there was an article about MySQL and ACID.

its all about the pricetag

[siliconwafer]

Although MySQL might have a huge market share, let us think about who uses it. Stating the obvious, most sites using MySQL now don't have the financial resources to purchase an Oracle license and wouldn't be using it in the first place. I'd say Postgres would be a much greater threat to Oracle, as it's quite clear that Postgres has a larger featureset than MySQL and is geared toward users who need those features.

Funny?

[hackrobat]

Ah, what a pity! Not a single comment rated Funny?! I didn't think the topic of databases was that boring. ;-)

Firebird is a far better choice

[OYAHHH]

I evaluated MySQL and found it to be a very immature product.

Someone here on /. mentioned that Interbase had an open-source equivalent called Firebird. I've been using it for 6 months now and I would never ever consider MySQL again.

I don't get the feeling that Firebird is up to Oracle capabilities and MySQL certainly isn't up to Firebird standards. That should be a hint.

It's really a shame that MySQL gets so much air-time while being so weak compared to even it's free competition.

Re:Firebird is a far better choice

[gwl]

For fun, download the latest Firebird
and see if you can get it to compile
under Solaris.

It might be a "shame" that MySQL "gets so much
airtime" but at least it compiles.

What do you mean by online backups?

[Anonymous Coward]

With MySQL you have to lock tables before a backup. But that is hardly a showstopper for most applications. Once they are locked (and something else... check out mysqlhotcopy) the data files can be copied right out from under the database server. Takes a second, maybe? I can't think of any system I use that would be hurt by such a delay. Financial? Heh. When did any financial transaction take less than 5 or so seconds, anyway? ATM, online, etc... Plus you could consider MySQL's query log a backup. Heck, just run them on another database all the time, or mirror the main db... tons of options, here. The other thing that keeps coming up on this board is integrity contraints and checks. There *are* such things as dynamic contraints, or contraints entailed by the application, not by the database. Anything the database can do in this way, the app can do as well. And since you generally have to inform a user of any problems of this sort... why not just put it in the app?

Re:What do you mean by online backups?

[germanbirdman]

OK, let me explain how an online backup works:

For example the MS Exchange DB or MS SQL:

The database is put into "backup" mode, this means that all transactions do NOT get written to the actual database file, which can then be backed up. During this time, the database is still fully functional though, because all changes during this time get written to "patch files". So it is fully operational, just a little slower as it first has to check the "patch" files before checking the database files. And Database backups are mighty fast - like 35 Gig/hour or so.
Once the actual database has been copied, the database file is put out of backup mode and the patch file is also written to the backup media.
This is at least how it works for the backup vendor which I work for.
We do the same thing for Oracle.

The database stays fully functional (online) while we perform the backup.

This is what I mean by online backup.

Re:What do you mean by online backups?

[germanbirdman]

OK, I did some research...

mysqlhotcopy I did not know about, but it only works for MyIsam tables.

For InnoDB tables there is a tool available:

InnoDB Hot Backup

http://www.innodb.com/hotbackup.html

What about PostgreSQL?

Re:Yeah, I've seen it

[Tokerat]

I have never once had a question about MySQL the docs did not answer, and I learned with MySQL from knowing absolutely nothing about databases. So, tell us, oh all knowing troll, what exactly are the MySQL docs missing?

Re:If MySQL was just a bit more user-friendly...

[happystink]

I agree 100%, I think a huge amount of MySQL catching on has to do with the online docs which are great. Most people getting into mysql are new to databases, and having all that info super-handy in a great structured format is very very handy, yup. I would have never got into mysql without it either, because for years there were no books on it.

Missing Features? - its getting there...

[Precipitous]

A variety have posters have complained about missing features. Some are already available - and the poster just stopped following MySQL at v 3.0. v 5.1 will have most of the features you'd expect in a relational database. Another post has already clarified that transaction support is available via Berkeley DB tables.

last I heard, views weren't on the list ...

Views will be a 5.1 feature. Check out MySQL Database Server features list for the the current development plan. (Search "upcoming features" to find their release plan.)

At any rate here's my favorite feature. I can open mysql and run a query summarizing all charges (from a 23 Million record table) by financial class for my boss in a certain time frame all in under a minute (query time: '13 rows in set (42.40 sec)'. Time to open mysql command prompt, and launch a query from a file:

Re:Missing Features? - its getting there...

[Ragica]

Good grief, MySQL 5.1?! I just scurried over to mysql.com to check this out. As I thought... 4.x isn't even released yet (the download page says 4.0 is "gamma", and 4.1 is "alpha"), and already the myfanatics are flaunting the 5.1 features?!

Based on this, I think the big boys should definitely be afraid of mysql. Mysql, it seems, is blowing them all away in the vapourware arena!

go MySQL! wait...

[mattwolfewvu]

(realizes that his school's CSEE department is named after a former COO of Oracle who donated a lot of money) Go Oracle!

Maybe also under LGPL?

[kcb93x]

Check...it might also be licensed under the LGPL, which is much like the BSD license...maybe that one too...Check it out. I know OpenOffice.org is licensed both GPL and LGPL.

Re:Have a look at SAP DB before talking about thos

[Aussie]

SAP DB [sapdb.org] is free, open source and GPL. It also has all the best big-guy features. Not many people seem to know about it - it certainly has small mind-share. But it is the real stuff - miles ahead of MySQL.

But for some reason people ignore it. Is it because it is created by a company and not a group ? Or is it that everybody has already chosen their favorite free DB and won't look at others ?

Sure beats me.

Re:Oh, wow, I am out of my league...

[mitcharoni]

Ah, spoiling the open source wet dream again. What a feeling!

When someone audits and publishes a TPC-C run using MySQL that's within 3% of the price/performance leader, I'll eat my left shoe.

I'm platform agnostic myself. Don't care either way. But bigotry for the sake of being a bigot is just stupid.

MySQL is appropriate, even for small stuff

[crucini]

I disagree with the idea that small projects should use flat files or XML in place of MySQL. First of all, the flat file only looks good while there seems to be a single entity in the system - let's say person. It rapidly turns into a convoluted mess when a second entity rears it's head - let's say a person can have multiple cars. Second, many applications end up developing reporting requirements that were not envisioned in the original design. That's what makes relational databses great - ad hoc reporting.

Another way to put it - as the application grows in complexity, more functionality will be added to the data store as the programmers painfully rediscover all the challenges which real databases have already conquered. Of course MySQL doesn't cover all of those, like ACID, but it covers most. Look at the amount of effort that went into MySQL, Postgres and Oracle - it's huge.

Of course, you may be thinking of simpler applications than I am. If the data can legitimately be represented by one table, with no denormalization, then I agree a database may be overkill.

Re:MySQL is appropriate, even for small stuff

[RevAaron]

Yes, I'm talking about some pretty simple applications with relatively small amoutns of data. Things like single-person todo lists or schedules. Over years, this data could amount to 1-2 MB maybe. All fitting in one or two tables. If you needed to add a new column, it would be pretty easy to do in an XML file or a serialized hash or array, and even easier to do in an object database.

Re:MySQL is appropriate, even for small stuff

[burnsanthony]

Im glad you said it so I didnt have to, the idea of doing everything with a text file is ridiculous

Re:MySQL is appropriate, even for small stuff

[Zaiff Urgulbunger]

If the data can legitimately be represented by one table, with no denormalization, then I agree a database may be overkill.

Agreed. I think there should be more of a drive to build a standardised DB-like API like MS ADO - this way you can use a simple flat file based system where you're dealing with small amounts of data and very few users, but (importantly) be able to scale up to a more industrial grade solution such as MySQL as needed.

For example, with ADO you can retrive data from an Excel worksheet much as you would from a database (I guess I could include Access here!).

I personally find this a very useful feature -- accepting that it isn't perfect and often you do actually need to make some changes!

Then again, I always find it a total pain in the arse not being able to use stored procedures unless the DB supports it, so working with Access is always a painful experience.

Re:If MySQL was just a bit more user-friendly...

[mitcharoni]

Based on that, I would presume then that MS Access could do 100% of what MySQL could do since it has none of those features either. So, let's just all "upgrade" to Access then. Hell, who needs data consistency and performance anyway?

Hmm I would be nice but.....

[snero3]

It would be nice to see MySql get up and take the big boys on (Oracle, DB2) but I don't think it is quite ready for that.

Speaking from an Oracle point of view, as I haven't worked with DB2, it just has too much going for it atm. For example the ability to do

create table2 as
select *
from table1;

Hey Presto you have an exact copying of table one(constraints and all). If you add a few more characters (not commands just characters) you can move it to different tablespaces, schemas and even whole databases on completely different machines. With oracle 9i you can move whole data files, turn on and off tablespace enlarge and reduce tablespaces all on the fly while the db is running. You can't imagine how much time this and 100's of other nifty features saves,

I am not knocking MySql, I think for websites like slashdot where speed/uptime seem to be the main factor and not table complexity/integrity/constraints it does great (how many times has slashdot been slashdotted?). However, having used MySql I don't think it is up the job of replacing a database like oracle.

On the other hand a couple of years go Linux was not ready to take on solaris, AIX etal... and now look at it? I suppose you never know what will happen!

ReiserFS

[alder]

It's too bad most Linux developers aren't interested in doing something really forward-thinking. If there was a DB integrated into the OS...

If you have not heard about it by now here is a useful link -- ReiserFS whitepaper. It is designed to do everything you mentioned and more.

Firebird

[puddytat]

Firebird should do what you want. It came from interbase so although Firebird 1.5 is still in beta it was allready a fully featured RDMS at the start. Big advantages for me are that it is completely free (beer&speech) and also supports windows as a server platform (which some of my clients prefer)

Re:Firebird

[Graspee_Leemoor]

I hated Interbase. It was god-awful. Anyway, you don't get that "open source zealot" recognition with Firebird. There's probably fewer people working on Firebird, less written about it, and its name doesn't inspire people to switch their company to it.

graspee

FireBird AKA InterBase is also interesting

[leonbrooks]

Have one compatriot using it with great success and singing its praises but have never seen a competitive analysis.

Not a DBA, but...

[alder]

Just an addition to your list: Firebird -- used to be InterBase.

PostgreSQL also runs under Solaris

[leonbrooks]

[M]ySQL also works quite well on Solaris

So does PostgreSQL. When they polled their users, four percent claimed to run it on Solaris (4th most popular after Linux, Windows and *BSD).

MySQL AB comments

[martenmickos]

Great discussions on this thread! We are reading them carefully to learn what we can do better.

Let me just comment on the overall impact of having such articles appear on Fortune.com and CNN.com:

The article is indeed the result of PR work done by MySQL AB, but the value of it will benefit the entire free software / open source community. We need to get many more business articles out there, so let's be happy about this one, and let's produce more of them!

Although this very article mentions MySQL only, please have a look at other articles where we at MySQL AB consistently mention the other open source databases. Here are two such articles on prominent business-focused sites (one of which, incidentally, is powered by MySQL):

http://www.open-mag.com/01943583279.htm

http://www.alwayson-network.com/comments.php?id= A2 44_0_1_0_C

Our ambition is not to be a threat to bigwigs per se, but to make superior database software available AND affordable to all. With your help we can do it.

Marten Mickos
CEO, MySQL AB

Welcome to Earth...

[leonbrooks]

Welcome to Slashdot, I sure hope you didn't expect fairness.

Welcome to Earth, I hope you didn't expect fairness.

I guess SlashDot is Real Life(tm) compatible.

Speaking as a Replicant...

[leonbrooks]

(-: 'S'true, one of the Replicants in Blade Runner was called Leon :-)

reliable replication (argue away, only shareplex, NT SQL server & Sybase have it today)

Not arguing, just pointing out that PostgreSQL replicates reliably [PDF] too. (-:

Re:PostgreSQL has every feature but Replication.

[Bradley]

Well, its all very well to claim speed, but its not always true. Yes, MyIASM is (generally) quicker than postgres.. The problem is that innodb is really slow.

For example, imagine that I have a bugs database (which I do; its called Bugzilla). I want to find all bugs where I'm either the reporter, or I've commented. Since bugs can have zero comments, I need to outer join the busg table to the accounts table. With randomly generated data, and 100,000 bugs, 1,000,000 comments randomly distributed among the comments:

select distinct bugs.bug_id FROM bugs LEFT JOIN longdescs USING(bug_id) WHERE (bugs.assigned_to=86 OR longdescs.who=86);

mysql: 3.33 sec
postgres: 14423.07 msec [I can get this down to 11758.74 msec by changing a
config option - I should probably file a bug on that]

(returning 46 rows)

But the thing is that I just want bug numbers. I don't care how many times I've commented, or need to know the contents of the comments. What I really wanted to ask was:

> SELECT bugs.bug_id FROM bugs WHERE bugs.assigned_to=86 OR bugs.bug_id
IN (SELECT bug_id FROM longdescs WHERE longdescs.who=86)

mysql: error
postgres: 137.66 msec

(To be fair, this does need CVS-postgres to be fast. It is, however, possible to create a UNION join to make it much faster, arround 6ms. Thats not appropriate for the generated SQL bugzilla uses, and having the db do the transformation from teh LEFT JOIN requires knowleges of the distinctiveness of bugs.bug_id, and the distinct in the original query, so its not trivial)

Oh, and it took almost 2 minutes with innodb. I will admit to not having tuned the various innodb parameters, plus this was 3.23, so I'm sort of ignoring that.

Queries are often slower, too - as far as I can see, mysql doesn't produce a query tree, but rather a list. If you only have one table, then this doesn't matter, but if you start doing complex joins (and without subselects, you do that more often than you may want to) then it does make a difference. Yes, I know that mysql has subselects in the development version. I'd be interested to see if they have the same speed problems as in postgresql I see it as one of the main advantages of MySQL over PostgreSQL is that you are able to turn off transactions if you don't need them.

You always need transactions. If you think that you don't, then its obviously not important to you if your data gets lost. (You also want foreign keys and constraints, but....) And whilst I don't know too many 'bigwigs', I'm pretty certain that most of them want half of their commited data to not vanish if there's a power failure.

Now, I'm certainly not saying that postgres is always better than mysql, and I'm sure that people can come up with similar examples to the above, but where postgres sucks and mysql blows it away.

Re:Sounds like a job for the Gimp!

[realkiwi]

> I had to do some work on a Postgres system over the weekend. I come from a MySQL background. ... and
>Having used both, I can honestly say that I'd rather wait for MySQL to add the one thing I'm actually waiting for: stored procedures, than use Postgres. Postgres just gives me the shits. Maybe that's why MySQL is the No1 open source database.

Excuse me but "had to do some work over the weekend" isn't the same thing as "having used both"...

You had the time in one weekend to learn all the stuff that most PostgreSQL admins learn over a few weeks or months?

I use PostgreSQL because it has always been free and it has never lost data. I haven't had a command line segfault since 1997 so I guess I must also have my server set up correctly...

And if I was to say bad things about MySQL (which I couldn't use in 1997 because it wasn't free) I would take at least two weeks to learn it - not just 2 days.

Its more to do with all software

[ExEleven]

Its not just all database software, its all software, I am a member of Linux Users Victoria and out our last monthly meeting a person from MySQL get to do a presentation, here is a quote "Although we would like to take over the world, we just like to make good software, and also, we cant really take over the world because oracle is very good at being oracle, and Microsoft are VERY good at being Microsoft." But this shows that the capitalist arguments against free software are a) wrong b) biased c) thought in denial.

Re:Have a look at SAP DB before talking about thos

[mentin]

MSDE (free version of MS SQL) is also free (as in beer), not open source, and has most the best big-guy features. Not many people seem to know about it. But it is the real stuff - hundred miles ahead of MySQL. :)

Egos & MS bashing

[tqft]

"Can you imagine the developers, who have big enough egos to include three pages of MS SQL Server bashing in their docs"

reading Slashdot. I can.

Maybe you should have used the AC option.

Can they fire you?

Re:If MySQL was just a bit more user-friendly...

[Ragica]

PostgreSQL has vast searchable "interactive documentation" as well.

MySQL bashers out there

[forged]

The article reminds me of this particular piece, somewhat old I agree (2000). Follow some extracts.

There are very good reasons for using MySQL. A need for a reliable, ACID-compliant datastore isn't one of them.

A Few More Details

• MySQL has no subqueries.
  Instead of performing one complex query that is entirely processed on the database end, MySQL users have to perform 2 or more serial queries that each must go over inter-process or network communication between the app and the database. This significantly reduces the speed advantages of MySQL.
• MySQL has no stored procedures.
  If a series of DB actions need to be performed in a block, MySQL requires each SQL statement to be sent from the app, again in a serial manner, again over IPC or network.
• MySQL has no triggers or foreign key constraints.
  Data invariants must be maintained by application-level code, which requires building carefully-planned abstractions to guarantee integrity (for every means of accessing your DB), and even more unnecessary back-and-forth communication between the app and the database.
• MySQL only has table-level locking.
  Only one user can write to a table at the same time. For web usage, that falls under the category of "pathetic."

The Bottom Line: MySQL is just a glorified filesystem with a SQL interface.

Summary of arguments for MySQL

[Ragica]

Having browsed this thread I notice there are certain common arguments frequently employed to defend mysql. But what kind of defense are they?

1. Its good enough for 90% of the web applications out there. However. 90% of the web applications out there are moronic, programmed by people without clues. This is a defense?
2. Similar to above: Most people have no need for the "advanced features" of ACID databases. However, most people actually don't realise what excellent use those tools frequently are, and with mysql they will never have a chance to learn them.
3. People don't always chose software purely on technical merit. One poster actually (apparently defending MySQL) held up microsoft as an example for this. I think this previous "defense" shall by the only comment I will make on this point.
4. Mysql is blazing fast. This has been debunked so many times, it is just tiresome to do so again. I'll merely rephrase another reply in this thread: sure, because it doesn't have to do half the things a real database does to ensure integrity, and functionality. Nearly any database server can be tuned for speed by disabling (or just not using) features, and proper index maintenance. Mysql on the other hand, can't be tuned for the features it is missing.
5. Mysql is easier to use. In my experience as a database hosting provider, pretty much the only reason this seems to be true is because unfortunately the majority of users got their feet wet in mysql first, and their brains apparently became irrevocably corrupted from the experience; they no longer are able to recognise the proper way of doing things even when it is shown them.

   None of the above is new information. Just my personal summary. In short, (ie. in troll) the argument seems to be simply: "Yes, mysql is pathetic, but so are most of us." Great.
   

Definitely in that case

[Quila]

Slashdot is running MySQL, and we've seen the load it can take.

Except

[Betcour]

When you drop the table, you'll also have to manually drop the sequence as well, which kinda defeat the whole purpose of using "serial" instead of creating a sequence.

Earlier than that

[Betcour]

MySQL 3.23 has foreign keys, transactions and "basic" triggers ("on delete/update cascade"). All you need to do is use the innodb table handler when creating your table.

MySQL vs PostgreSQL vs MS SQL Server

[jdoeii]

The following is my personal opinion based on experience with MySQL 3.x, PgSQL 7.x and MS SQL 6.x-2000. YMMV.

-- MySQL
MySQL's #1 feature is exceptionally simple use and administration. #2 is excellent read performance. If you are building a web site with a simple DB backend, a lot of reads, a few updates/deletes, no transactions, then MySQL is ideal. It may have an infrequent index corruption problem. A couple of versions had a problem of eating all CPU on occasion. But most of the time it just runs. Very easy to learn, nearly 0 maintanence.

The problems start when the database grows in size and complexity. DISTINCT sucks: if a query is large enough distinct just does not work. UPDATE/DELETE don't work with joins. No transactions. Yes, yes, I know, some things are addressed with another table format, some are patched up in 4.0. But it's done at the cost of what I listed above as #1 and #2.

-- PgSQL
PgSQL is the #1 OS ACID-compliant DBMS. It has transactions out of the box. It works. It can be tuned to run as fast as MySQL. But the learning curve is steeper both for programmers and for DB admins. The first thing to learn is VACUUM ANALYSE. If the table has a lot of updates/deletes, do it often, otherwise performance would suck. And I mean really suck. When the the V/A is running, the table is pretty much unaccessible. Pg's aggregates are bizzare. MIN() and MAX() can be programmed around, but if you need COUNT() you are out of luck. Flat out can't use it on large tables because aggregates do not use indexes! Another surprize is difference in performance between queries WHERE ... IN () and EXISTS. The difference is orders of magnitude.

Pg's query planner/optimizer is not too smart. It can be easily confused with variable types. It often chooses wrong indexes. It's about the same quality as the p/o in MS SQL 6.5. Be prepared to spend time tweaking the queries and indexes. On the other hand, the memory cache is too smart. You can't make it read the whole DB into RAM and keep it there. Pg is trying to behave nicely and releases RAM when the table is not used for a while. Which is a bad thing on a dedicated server.

Hopefully, the next release will have failover and replication. And maybe aggregates will be eventually fixed. Then PgSQL can be seriously considered for enterprise-scale projects.

-- MS SQL
Microsoft went a long way from MSSQL 6.5 to 2000. 7.0 and 2000 are pretty much the same. If it were an OS project, 2000 would have been called 7.1 or 7.2. The MSSQL 2000 is a pretty solid product. For an all-Windows shop it works. But the price! The MS SQL server license, The NT Server license, a license for each connection. If you need to access it from a FreeBSD or Linux box, then FreeTDS is needed. FreeTDS has its own issues. MS SQL has a problem with scalability because it's tied up to NT and NT is not available (yet) on massive hardware.

In conclusion I would say that MySQL is certainly no threat to big boys, not now, not in the near future. PgSQL may become a real contender in the next two years. It is already eating into the lower end of the MS SQL market. Unless MS ports the SQL Server to other platforms, it's going to be sqeezed really hard.

Re:BUSH = JEW/NIGGER LOVER. FUCK SQL

[tha_mink]

1). Colin Powell -- Secretary of State, ex-chairman of Joint Chiefs of Staff, of Jamaican origin, he has one Jewish ancestor on his father's side. He grew up in a heavily Jewish-populated neighborhood in New York, and speaks Yiddish. ..." A heavily Jewish-populated neighborhood of New YorK???? Have you ever BEEN to New York. That's like saying a heavily cornfed-populated neighborhood of Iowa. Dumbfucker...

DB2 is integrated into OS/400

[emil]

If IBM found compelling reasons to do so, then said compelling reasons must exist. Now, you don't see AS/400s at the top of the charts at tpc.org, so I would tend to believe that it isn't necessary for high performance, but that doesn't mean that there is no benefit at all.

p.s. With no views, no triggers, and no foreign keys, I wonder why MySQL gets all of this hype. There are so many free databases that implement these features and don't seem to suffer from performance issues.

Re:DB2 is integrated into OS/400

[Arethan]

From a technology standpoint, MySQL is indeed quite lacking. However, from a usability standpoint, MySQL is absolutlely fabulous. The installation is incredibly easy, whether you use the prebuilt static binaries, compile from source, or use the packages that come with your Linux distro. And using the application is simple as well. Not to mention that the documentation for MySQL is superb.

But in the end, you are correct. MySQL doesn't scale to enterprise levels because of the lacking feature set. It is perfect for websites and as a storage medium for certain applications, and that is really about it. There's nothing wrong with being a niche product, and MySQL fits its niche perfectly.

Are you using the free Sybase?

[emil]

Sybase ASE version 11.0.3.3 is free for production use on Linux.

Since this is my (tax) money that we are talking about, please tell me that you are using this version for non-critical applications.

Re:Are you using the free Sybase?

[dasmegabyte]

They consider it non critical. It's a database of property tax exemptions. The tax database is a db2 monster.

And linux at this level of government is silly. No training budgets and most IT departments are one guy who's, like, totally good with computers, who comes in twice a week to uninstall KaZaa and VirtualGirl.

Check IRL results

[leonbrooks]

Among the enterprise-level DBs, SQL Server beats the crap outta every other DB both in price/performance, price AND performance.

That depends on who classes what as `enterprise-level'. PostgreSQL eats SQL Server's lunch pretty much across the board except on carefully architected benchmarking systems (think MindCraft and four separate logging partitions). SQL Server is basically only warmed-over SyBase, even though a lot of the later warming-over was supervised by a very competent ex-DEC database bloke.

OBTW, don't trust Microsoft's costings for their competitor's systems as far as you can throw them.

For example, they sponsored a recent IDC study which assumed a software cost of USD$940/1009/340/6609/1390 (networking/file/print/security/web) for each Linux system (real cost $0/0/0/0/0, versus (they claim) $211/$3988/1665/5829/7107 for Windows 2000) - which factor alone made more than a slight difference to the resulting TCO figures. Microsoft-sponsored-IDC quote staffing costs for Linux at significantly higher than those for Windows, where in real life the reduction in headcount brought about by increased reliability prunes those costs to about 1/2 or 1/3. They also quote downtime costs for a Linux-based file server at over $4000 a year and nearly double that for a print server! I could do a complete reinstall from scratch every quarter and still cost less than that.

Being pessimistic and assuming that a Linux server collapses half as often as Windows, and nett staff costs are around half those of Windows, the bottom lines for Linux work out to: Network 64% File 54% Print 55% Security 56% Web 40% of Windows. Quite different to IDC's own conclusions.

IDC also do $3 worth of outsourcing for their WIndows file server. Good trick.

Re:Check IRL results

[melted]

The thing is, these results (I mean TPC) are provided by HARDWARE MANUFACTURERS, not Microsoft. It's not like there's a bunch of people in Redmond sitting and faking these numbers. And it's pretty legally dangerous to fake them.

Amen

[abulafia]

Now that I have tried Postgres though there is no way I would go back.
I grew up on Sybase and Oracle, and couldn't stand the constraints of Msql/Mysql when I first started using them.

Now that I'm using Postgres wherever I have a choice, I even prefer it to Oracle.

About the only problem I have now is that I support too many different databases, and switching between ';" and "go", and "show tables" and "\dt" drives me nuts.

In any case, Postres rocks.

"I think"

[MosesJones]

So please, tell us how many different databases you have DBA'ed on or architected. What functionality the "big guys" have that MySQL doesn't have.

Oh wait, you have never done anything beyond a website ?

Couple of hints: MySQL and OLAP, MySQL and Clustering, MySQL and XA compliance, MySQL and hot backups, MySQL and schema migration etc etc etc Sure it has _some_ of this stuff. But can it compete head to head with the "big guys".

The idea that MySQL will replace Oracle across the enterprise is laughable, the support agreements aren't there, the reliability and history isn't there. And the last comment sums it up.

Have you _any_ idea how much people apy for those large data bases and data warehouses (now MySQL for Data Warehousing.. please) ? Hint: It makes Web Sites look like chicken feed.

Enterprise systems is worth more as a market than the Web Systems. And MySQL does _not_ have the facilities to compete. What it would need the support of SAP, Baan, JDE, Peoplesoft, Siebel et al, as well as a whole lot more functionality.

People Web != Enterprise and $Web $Enterprise

Re:"I think"

[lewp]

ERROR 1064: You have an error in your SQL syntax near 'People Web != Enterprise and $Web $Enterprise' at line 1

Re:Have a look at SAP DB before talking about thos

[LazySlacker]

I believe it has a limit on the number of concurrent users (5?).
Also getting hold of it is a problem, well it was for me the 2K/XP compatible version doesn't appear to be on the web site I think you need the pro version of the OS :-(

definitetly not adequate for 99%

[clarkc3]

no relations between tables, no stored procedures - sorry, never gonna hurt the Oracle\DB2\Sybase's out there for the hardcore database stuff til they do something about that - which they wont since not having those is what makes it fast at other things (they even state that in the docutmentation)

Re:definitetly not adequate for 99%

[letxa2000]

You, and someone else higher in the thread, seem to have completely misunderstood my post. I'm not saying that MySQL is going to take the enterprise DB market. That's not the case.

What I'm saying is that 90%--maybe even 99%--of all database installations do NOT require the features offered by some of the big boys. I would go so far as to say that many applications that are using "big boy" databases could get by quite well on MySQL. I know of at least several cases where companies are dumping MS SQL Server to go for a web-based, Linux/MySQL installation. For them that makes sense and I'd bet it makes sense for many other companies.

The point is not that the "big guys" don't have a place. They do. The question is whether or not they can remain profitable when those applications that don't NEED all their features go to a free database (MySQL) that gives them the features they need rather than paying big bucks for "big boys" that give them more than they need. A multi-national company with DB thousands of DB clients all over the world probably needs a "big boy." A company with a hundred DB clients most likely can do just fine with MySQL and can do it cheaper than going with the big boys...

That's all I'm saying. Don't get all defensive!

FOSS Reports

[leonbrooks]

From FreshMeat:

• Agata (report gen, PHP-Gtk)
  
• DataVision (report gen, Java)
  
• SQL Databases for Linux (article, old but lotsa links)
  
• DataMiner (mining/editing frontend, PHP)
  
• DBA-Dialog (DBA tool, BASH (yes really!))
  
• DB Designer (complex-SQL-database design tool, apparently C++/Qt)
  
• Foxxess (DBA frontend, apparently C++/Fox)
  
• knoda (define, populate, report on DB's, apparently C++/Qt)
  
• Koala (similar, Python/Gtk)

Lots more out there, I'm sure... you can easily skip some when your brain melts down after scanning 200 entries and maybe clicking on 50 to see if they're as relevant as they seem. Still much easier than reinventing the wheel yet again.

Re:Have a look at SAP DB before talking about thos

[frostman]

I have an Oracle background, and for a new project (my own) I decided to use something I could afford (ie, free). So I checked out SAPDB, becase here in Europe SAP has a certain blue-chippish reputation.

I have to say, I was pretty disappointed, not in the DB itself but in the documentation, installation, supporting programs, and so on.

Simply put, it's a MESS. I wouldn't be surprised if it runs well (I didn't do any performance testing), but I sort of got the feeling the Big Companies using it probably had their SAP consultants do all the work, and never had to touch it themselves.

I ended up with Postgres, which I'm quite happy with. Excellent documentation, good online support forums, decent if not great admin tools for Windows, they don't pretend that the DB works on Windows (though some company makes a supported Windows version). Oh yeah, and slowly, slowly some of the better database design tools support it. And I hear it's got a great admin interface for Linux, which I hope to try out some time.

There's even a phpMyAdmin-style web admin tool too if you're into that.

I understand why SAP doesn't want to pour anymore cash onto the product, but as long as the community is so small, I'm afraid people will continue to ignore SAPDB or, like I did, evaluate it and move on.

Even if it's a really good DBMS as such.

No, MySQL is not a threat.

[vorwerk]

MySQL works in a completely different "realm", if you will, than Oracle, DB2, Sybase ASE or ASA, etc.

Reasoning:

MySQL is a great replacement for a file system-like Berkeley DB interface. Works fine for small jobs, but it isn't ANSI SQL compliant, and doesn't necessarily support the same level of features and robustness (transaction-level rollback/commit/etc) as do the major vendors.

A prof once told me: When you buy a license from one of the big DB vendors, you aren't necessarily paying for their performance or features. (They're at a point now where they all run pretty well the same for the average user and offer pretty well the same features.) What you *are* paying for is the "guarantee" of stability and robustness in the event of a critical failure.

That said, I use MySQL for my webserver, and Sybase Adaptive Server Anywhere for my business operations. ASA is fast, cheap, incredibly easy to administer, and I know that I can rely on it for my data integrity. IMO, the same cannot be said (with such confidence) about MySQL, and therein lies the difference.

Re:Have a look at SAP DB before talking about thos

[RoundSparrow]

I agree.

We have been using SAPDB for over 2 years, and it is very disorganized. There is no replication of _any kind_ (Microsoft has this with SQL Server and so do others).

I like that it is ivery CLI driven, but the code is a mess of Pascal converted to C and the drivers are poor quality, etc, etc.

What about Windows 2000 Indexing Service?

[Ride-My-Rocket]

Granted, it operates solely against files for the time being, but it does create indexes for faster searching and such. And it does make metadata (user-definable or auto-extracted) available for searches...........

Re:You meant to say "pre-relational"

[javabandit]

You meant to say "pre-relational." Pick and other multivalues databases have been around for a long time, sure. That makes them "pre-relational" because they predate the big shift to relational databases.

No. I mean to say "post-relational". Because they support 1NF as well as multi-valued relationships. Which means that they handle more than 1NF.

The very concept of multivalued columns is a violation of 1st normal form, though, so they are not on a sound theoretical foundation the way RDBMSs are.

Multi-valued databases support 1NF as well as multi-valued relationships. If someone wants 1NF, they can use a multi-valued database for that. MV databases have full integrity constraint support, PK columns, the whole nine. They are there for the taking.

Where MV databases shine is in the denormalization of data. 1NF by itself simply doesn't cut it (performance-wise) for these kinds of functions. Which is why Pick-based/MV databases are used by huge financial institutions and ERP vendors.

What amazes me is that a lot of people think Oracle is actually the first one to tackle this problem with their materialized views and such. Some think star schemas were first to deal with data crunching, et al.

MV databases have been doing this for decades. And still today, they do it faster than any of these databases do it.

Features / cost not the issue

[walt-sjc]

MySQL is great for lot's of tasks, but adding subselects and transactions isn't the issue. Postgres has Loads more features than MySQL, yet it isn't really replacing Oracle in the enterprise. Why?

It's the applications support. If you are buying a $5M ERP, CRM or whatever, it's going to support Oracle, and possibly DB2 or MS/SQL. In order to "port" to another DB, it's going to take a Huge effort with vendor support.

When third party commercial developers start supporting open source DB's, THEN you will see enterprise adoption. In order for commercial developers to start supporting OS DB's, they need to see enterprise demand. Before you will see enterprises requesting MySQL support, MySQL needs to have the feature set needed to support those applications (such as triggers, stored proceedures, etc.) and enterprise level support for clustering, load balancing, replication, etc. MySQL is beginning to get some of these features, but it's not NEARLY at the same level as Oracle.

It's the same situation as it is with operating systems. If the applications you need for your business are only available for the Windows platform, what real choice do you have?

RTFW! (Read The Fine Website)

[jeremiahstanley]

You cannot make SQL queries against XML or flat files. Integrating a DB into the OS is a good idea, that is why we have filesystems. But, v4 of MySQL has a library that can be embedded into applications that will let them make their own MySQL like data stores that can be accessed with a SQL query. You do not need to have a database server running, it is in the app. You do not need a database for your todo list or your contacts, that is just bad app writing (unless the app is being written for 1000+ users (LDAP!)).

Re:RTFW! (Read The Fine Website)

[geekoid]

" You cannot make SQL queries against XML or flat files."

true, but I can use grep just as easily as SQL.

Re:RTFW! (Read The Fine Website)

[jeremiahstanley]

I've never seen grep do an outer join of tables nor stored procedures to maintain relational consistency.

Re:Bull.

[SatanicPuppy]

MSSQL is worlds fastest database AND supports true ACID compliance

Man, you just made me snort milk through my nose.

MSSQL goes down more often than a 50 cent whore. It is completely undocumented. Everything is wizard driven through wizards that tend not to work. It is a huge memory hog, and it costs WAY too much for what it offers.

Comparing MSSQL to Oracle is a joke.

My 0.259338 Austrian schillings(obsolete) worth.

Re: a select...shouldn't lock the table

[sqlgeek]

With respect to "block and row level locking. A select on a 50 million row table shouldn't lock the table." I can only say that a read (i.e. select) should never lock a table in anything approximating a modern database. I think that you meant that a write (insert/update/delete) should not lock rows of a table that are not targets of the statement, yes?

Scott

Re:Have a look at SAP DB before talking about thos

[julesh]

But for some reason people ignore it. Is it because it is created by a company and not a group ?

Errr... MySQL was written by a company as well. TCX DataKonsulter AB are the company of IT consultants who originally wrote MySQL.

is MySQL the right tool for the job?

[juan2074]

For a lot of things, MySQL is more than enough.

I think the original poster meant that MySQL is being used for many things that really do not need a DB. If something can be done efficiently and effectively in a plain-text flat file, why not? Sometimes, developers find a reason to use MySQL when it is more than really necessary to accomplish the goals of a project.

Use the right tool for the job. MySQL is probably overused right now.

Firebird Flies

[AWrinkler]

We use Firebird in Enterprise OLAP capacities for a few big clients.

0.5TB Single File Databases, High speed built-in stored procedure language, simple UDF (User Defined Function) interface for funky stuff written in C, Delphi, PERL, whatever.

We've never had a requirement for MS SSQL Server, Oracle, DB2, Empress, PICK, or Postgres.

Firebird just does it all, and blazingly quick at that

The name on the shingle

[leonbrooks]

these results (I mean TPC) are provided by HARDWARE MANUFACTURERS

And so...? The TCO study was provided by IDC, using carefully chosen figures as expressed by Microsoft when they requested the study. The name on the shi8ngle doesn't matter nearly as much as who actually provided the impetus.

Think back to MindCraft, where they did things like using 4x100Mb cards instead of 1x1000Mb like anyone else would, and spreading Windows' logging across four partitions. Sure, they finally managed to tweak good results out of it (and of course the Linux kernel was so much better in those areas three months later anyway), but who in their right mind is going to set up a production machine like that, or even know that that doing so would make it faster?

Likewise choosing high-end hardware specifically for your TPC and tuning it to within an inch of it's life is not what your average punter is going to do, and only vaguely related to what a good DBA will do. They're going to buy a stock machine, do a stock install, make a few of the more popular tweaks, and run that. In those circumstances, PostgreSQL will comprehensively pound MS SQL Server into the pavement.

Re:The name on the shingle

[melted]

That's why some system engineers and DBAs get paid real well. We're not talking about "average" systems here. We're talking about enterprise databases. If you don't know how to set them up properly, you better hire someone who does. I'd really be interested in seeing another independent test between the latest release of RH Linux (tweaked and patched ad nauseam) and Windows Server 2003 with IIS6. It seems to me so far that Linux is up for serious buttkicking again.

Re:The name on the shingle

[leonbrooks]

I'd really be interested in seeing another independent test between the latest release of RH Linux (tweaked and patched ad nauseam) and Windows Server 2003 with IIS6. It seems to me so far that Linux is up for serious buttkicking again.

For that to be fair, you'd want to use the latest Linux as well as the latest Windows (that is to say, if you tweak one you should tweak both properly). In which case there is no question but that Windows + IIS will get their collective butts kicked all over the stadium by Linux +Tux + Apache. There are no more rabbits for Microsoft to pull from their corporate hats in this domain, but Linux whupped them good last time I looked (some dizzying number of requests per second on a single-CPU P2-500), and is only just getting up to stride.

Re:The name on the shingle

[melted]

Who the heck needs a single P2-500 junker serving static content? Let's see some serious stuff. Quads, 8-way servers, clusters, buttloads of RAM and data, dynamic pages served by each side's engine of choice. Something remotely resembling reality of a big corporate customer and letting the OS realize its full potential. Last time I checked Windows ripped Linux apart on tests like these. And IIS 6 (surprise!) now includes a kernel module as well, so TUX has lost its only advantage. And when we consider ASP.NET for generating the dynamic content we pretty much see that in terms of technology and performance the "inept" MS is far ahead. Perhaps it's time for linux hackers to stop writing widget libraries and making screenshots and show the world what they're made of.

MySQL IS suitable for HEAVY use...

[Anonymous Coward]

Let me just say that we have been using MySQL as our front-end database for just about 2 years now. We replaced MS SQL 7.0.

For those saying that MySQL isn't suitable for high traffic web sites should know this...

We provide web servervices to many of the large Real Estate Brokerages across the US. We are serving up about 60,000,000 page views monthly. The MySQL server powering those page views gets over 350,000,000 questions monthly. Our primary database is sitting at around 15,000,000 records. Anyone considering a new database should not believe what other say and test MySQL yourself. I can't stress this enough! We did this with MS SQL 7/2000 and MySQL. MySQL met all of our needs and crushed MS SQL in all of our testing.

Needless to say, MySQL is VERY powerful in capable hands!

Also, hot backups are possible. We backup about 24G worth of data in just a few seconds with no degraded services!

You wanna see something *really* scary?

[leonbrooks]

Who the heck needs a single P2-500 junker serving static content?

More than 90% of the websites in existence. (-:

Not that a Tux-based webserver is limited to static content, it's just that the static content whips out the door fast enough to make the cables bulge. And if it can do that on a clunker, what can it do on a real server?

Quads, 8-way servers, clusters, buttloads of RAM and data, dynamic pages served by each side's engine of choice.

Mmmm. Single machines, do we go for the high-end balanced-and-tuned zSeries or one of SGI's monsters? Clusters... I don't suppose you noticed that the world's most popular search engine is a passel of Linux clusters? I vote for the 512-way server.

Last time I checked Windows ripped Linux apart on tests like these.

Only on tests where the majority of skull-sweat went into selecting a setup that favoured every one of Windows' strengths and suffered from every one of Linux's weaknesses. On a level playing field, or to put it another way, something `resembling reality', IIS's ass is grass.

And IIS 6 (surprise!) now includes a kernel module as well, so TUX has lost its only advantage.

Er, no? IIS has always been Ring 0, which is why vulnerabilities in it have always been a catastrophe. That hamster has already been whipped to death.

And when we consider ASP.NET for generating the dynamic content we pretty much see that in terms of technology and performance the "inept" MS is far ahead.

Catching up with Java (Real Soon Now) is supposed to be `far ahead'? Gimme a break!

Speaking of buttloads of RAM, that's the way to make Samba really party. It uses about 20-30% more RAM per instance for file sharing than Windows 2000, but OTOH can serve about 10x as many connections on the same hardware. So which would you rather do, pay for ten servers, ten sets of licences and figure out how to integrate them? Or pay for one server and double the RAM?

Re:You wanna see something *really* scary?

[melted]

Sorry man, but you don't know JACK.

1. Remember, we're talking about corporate customers who are interested in running large sites with strong database backend. Static content just isn't enough here.

2. I suppose you heard that the worlds largest webmail site runs on a cluster of Windows 2000 machines. :0) But that's not the point. The point is that corporate customer won't be willing to invest as much into clustering as Google did. They need high performance, low cost and READILY AVAILABLE solutions.

3. As far as I remember a lot of skull sweat from RH gurus went into that Linux machine as well in mindcraft tests. Did it help? A lot of skull sweat goes into oracle clustered benchmarks for TPC-C, and it doesn't help oracle either. Windows solutions are still twice as cheap and twice as fast at the same time.

4. IIS has never been Ring 0. Nuff said. :0)

5. Catching up in what? .NET is far ahead of Java in terms of speed, XML support, security, remoting, UI libraries, you name it! You've obviously haven't seen .NET technologies in person. It's amazing how much quality stuff they've cranked out in just 4.5 years.

Re:You wanna see something *really* scary?

[leonbrooks]

Static content just isn't enough here.

Y'don't get it at all... Tux fields the static stuff and handballs the dynamic stuff. Even on heavily dynamic web pages, many or most of the files and bytes are static (images, static frames and such) and that gets processed roughly an order of magnitude faster than IIS. Even the stuff the Tux handballs (to Apache, for example) can be processed damn fast (e.g. Zend scripting).

I suppose you heard that the worlds largest webmail site runs on a cluster of Windows 2000 machines.

Yup. More than twice as many machines as the same load needed under FreeBSD, and each more powerful, and it's never been as stable since, despite the addition of hardware load balancing.

As far as I remember a lot of skull sweat from RH gurus went into that Linux machine as well in mindcraft tests. Did it help?

The skull sweat went into making Linux run a test carefully designed to make Windows shine (four network cards, my ass! what kind of nutcase...? never mind), and it did help a fair bit on the day. But where it really helped is that Linux's performance in areas formerly ruled by Windows went through the roof. In particular the handling of multiple network streams to the same source improved, again, more than an order of magnitude within 3 months, and the `select' mechanism underwent the revamp that allowed the old clunker you so despise to reliably handle an unprecedented number of hits. Error 500s? We don' need no steenking Error 500s!

IIS has never been Ring 0.

Yeah? Look closer. And explain away its layer-violating `TCP cheating' (if you can) while you're there.

You've obviously haven't seen .NET technologies in person.

Oooh, yes I have. (-:

You've obviously only ever listened to the good side.

Re:You wanna see something *really* scary?

[melted]

And where are the benchmarks proving your point of view? Even those done on a machine carefully crafted to let linux shine?

Re:Offtopic

[Inthewire]

The trouble with that is drug tests don't look for raw drugs, they look for metabolites of drugs.br />Maybe straight coke would show something, but it isn't really the molecules they're testing for.
Good thought, though.

Mid-level datawarehousing and DSS w. MySQL

[jpa5n]

As people have pointed out in the various threads, folks get MS-SQL licenses for all sorts of things because they either already have a few licenses or because their consultants tell them they need them. Of course you need the MS-SQL license (or Oracle or Sybase or DB/2) for commercial ERP/CRM systems and some other hefty applications.

But what about decision support systems and (smaller) datawarehouses? Why should I license 4 processors of MS-SQL (e.g. 2 active, 2 for failover) for the DSS when I can get use MySQL? DSS/DW is mainly reading data -- stored procs, etc aren't that important, nor are transactions. Plus I can load MySQL with the batch update that is already using text files dumped out of the production MS-SQL server -- bcp on MS-SQL is a common way to load data. There's no good reason to drop another $40k on the reporting server.

I can even keep using Crystal Reports or ASP or whatever I'm already using. Plus I can run it on my existing Win2k servers (that's a strike against Postgres for now) without fighting over linux, etc.

Benchmarks

[leonbrooks]

Published to the LKML. Sorry, off out the door to RAID a machine in Belmont.

Oracle's feeling the sting

[droolinggeezer]

Depressing sales reported today, lower stock price. Guess more and more people are getting tired of those outrageous licensing demands. Guess Larry will have to use that same jet for another year...sigh.

Re:SELECT success FROM propoganda_status=

[ihatewinXP]

When I first saw this story come up I was wondering what you would have to say about it.... How can SQL's finest hour be your darkest? Keep your chin up, troll on,

flat files

[SHEENmaster]

Contain both configuration and documentation.