Slashdot Mirror

1. I wanted to know and also decide what is where. Automatic installation efficiently takes away your decision of whether Apache should remain in /usr/local/apache, /opt/services/apache or /home/apache (which I happen to prefer) or whatever you like.
2. I wanted to have the decision about do I need X to run nethack, or that my PHP module doesn't support MySQL at all but PostgreSQL instead. With a prebuilt package you might find out that to install say vim to your 486 with 280MB hard disk, you need to install X because the prebuilt vim happens to be linked against X libraries.
3. After I install any library (version) not known by the distro-of-choice my self, I efficiently render the dependency stuff broken so I end up compile it yourself system anyway.
4. Automated updatings are basicly nogood. What if my computer had updated to 2.4.15 as soon as it was released. I wound't mind that on a test machine but for my primary workstation I prefer having somebody else finding the worst problems first. Also IF the system breaks after update done manually, at least I can say what actually was updated. You need to use the update but if you don't, it's of no value.

The optimized for specific hardware thing might be an issue for a 386 which you want to run as fast as possible but really doesn't matter that much on modern hardware on which you often can't even find much difference with a program compiled with or without optimizations.

The great parts of compiling something yourself are so great because of the lack of automation, not because the compilation itself.

How does PostgreSQL compare to Oracle? Is PostgreSQL more or less secure than Oracle? I don't know. I've never heard of a problem with it nor have I had one. Is PostgreSQL faster or slower than Oracle? I don't know, and apparently Oracle desperately doesn't want anyone to find out. From benchmarks that have had Oracle results deleted to benchmarks that someone (I wonder who?) has gotten the ISP to remove for "violation of our Terms of Service" (this used to be a benchmark), Oracle is very aggressive in preventing anyone from finding out how their database really performs. I wonder why? (However what might be another version of the second benchmark seems to have survived by carefully avoiding the mention of names of proprietary products.) All I know is that after trying to deal with the bloat of Oracle on a less-than-mainframe-class PC, PostgreSQL was a lean, mean breath of fresh air. Converting PL/SQL to PL/pgSQL was easy, too.

The subject says it all... :)

PostgreSQL

There are several migration guides at techdocs.postgresql.org. There are none listed for Sybase, which makes me suspect there may not currently be one written. However, the other guides may have some good general tips for you.

One possible solution I came up with was to UUEncode the data, then store it in a text field.

I never actually implemented this, but it should be fairly easy to do (either outside the DB or using a Postgres user-defined function.

The existing BLOB support is documented briefly Here, but quite honestly it... well sucks.

Large object documentation

Or you can use bytea or TEXT like I do.

-Dave

Did you try the postgresql docs?

FreeBSD, OpenBSD

The options SYSVSHM and SYSVSEM need to be enabled when the kernel is compiled. (They are by default.) The maximum size of shared memory is determined by the option SHMMAXPGS (in pages). The following shows an example of how to set the various parameters:

options SYSVSHM
options SHMMAXPGS=4096
options SHMSEG=256
options SYSVSEM
options SEMMNI=256
options SEMMNS=512
options SEMMNU=256
options SEMMAP=256

No. Mac may have some "security through rarity", but OS X is not obscured. Neither are its web services nor its SQL implementations.

I've done some development with Oracle on Linux and quite a bit on Solaris. Linux seems just as capable an environment for Oracle. I've also spoken to people who have experience scaling Oracle on Linux and have gotten good reviews.

For much more, check out the oracle-on-linux mailing list. And here's a page at Oracle that says Linux is supported "across the entire technology stack".

But I don't want to plug Oracle on Linux too hard. You (for 95% of "you") probably don't really need to pay Ellison's sky high licensing fees. PostgreSQL really is quite good now and more than sufficient for most database applications, most of which barely scratch the surface of Oracle's scalability and capabilities.

I mention this, because it is something I find myself wanting to do all the time, for example, when storing data that originates in OO programs. Being able to store it in an RDBMS has heaps of advantages for me ... but I can't easily store the different info of different derived classes.

Speaking of DBMSes, one of the electrical engineers at work wanted to learn how to use the Oracle DB one of our projects is built on. Somebody told him "No problem, it's dead easy. Almost everything can be done with only two commands, SELECT and UPDATE. Simply learn those two and you'll know everything there is to know about DBs." Apparently, the CS guru who witnessed this nearly imploded. Wish I'd seen it...

I'm rather ignorant on this subject, but surely XML data could be viewed as being object oriented? If this is the case, then surely an OODBMS or more practically an Object-Relational DBMS could be used.

In case you were wondering if there are any out there, check out PostgreSQL which is way cooler than MySQL (it's open-source for a start)

I have worked on such a kind of universal contact database. I've implemented it several times as SQL tables and am currently working on an object-relational version (for deployment on PostgreSQL). I intend to release the SQL under the GPL when I'm satified that it is complete and correct. In the mean time if you'd like a copy at the earliest release opportunity please email me and I'll add you to a list of people who want early copies.

Allow me to describe briefly how my database is organized.

The database models people, companies, etc. as "contactable entities" (CEs) and allows any CE to be a member of any other CE for the purpose of arbitrary grouping. For instance if there is a CE for a company and multiple CEs for the employees of that company, the employees could have a membership relationship with the company CE. There is a parent CE relation that all specific CE inherit from (this is an object-relational feature of postgresql). It is possible to create new CE types by simply creating a new table and inheriting the parent CE class. Frankly the design has worked in the past with just people and "groups"

CEs can have any number of arbitrarily defined identifiers. For instance, some people want to identify their contacts by their full name, however, in some cases people are given unique usernames, or customer numbers, or other unique identifier criteria. These identifiers can be stored easily and not all CEs have to share the same unique identifiers. Currently I'm having some problem enforcing integrity constraints on these identifiers effeciently.

The database also models "contact methods" (CMs) in an flexible way. CMs are phone numbers, email addresses, phone numbers, etc. Just like the CEs, CMs are defined by sub-classing a parent CM relation/table. You can create new CMs but just defining a new table that inherits the parent. A CE can have any number of CMs and two different CE's can share one CM.

All CMs have validity ranges. So change of address actions can occure naturally. You don't just replace the old address with the new one, you actually keep the old address in the database, insert a new address record and mark the valid_end_date of the old one appropriately and the valid_start_date of the new one appropriately. This is very handy.

I'm currently working on adding tables for tracking actually contact instances (phone calls, emails etc) but have not completed any signifcant work on this portion.

In the past I've implemented this design about 4 times in SQL and each time it has been a little different with a lot of unique compromises. I'm working on a design that I hope with be more universal and truely reusable now.

"support. (who gives 24/7 support on postgress, and send out tech support guys giving consultations, will come on site on a sunday at 4am?)"

RedHat either already does or will soon.

"what OpenSource rdbms provide true mutli language support (we have records in cryllic, japanese, american, german, etc)?"

PostgreSQL.

"high availablity (i dont know the current state of HA functionality in the linux kernel)"

Why not look it up?

There's a good article here called "Why Not MySQL?" by Ben Adida [mailto], part of the OpenACS Project [openacs.org], on why MySQL wasn't the right choice for OpenACS (at the time). It's quite out of date (and is recognised as such by the author), but still worth a read, and there many interesting submitted comments. Take a look at some decent free RDBMS alternatives such as Firebird (open-source free Interbase) or PostgreSQL while you're at it. Oh, and there's plenty more dicussion on MySQL in a previous Slashdot article here.

Stef

In other news, PostgreSQL announced that it has all the features that MySQL 4.0 has, has had them for a long time, is GPL'd, and is very stable (as in not alpha). Get it here

OK, maybe that wasn't announced, but it should have been.

In other news, PostgreSQL announced that it has all the features that MySQL 4.0 has, has had them for a long time, is GPL'd, and is very stable (as in not alpha). Get it here

OK, maybe that wasn't announced, but it should have been.

Still no proper transactions, no subselects, no foreign keys or views. How can this be a "mission critical" SQL database? I still don't get why people use it. Sure, for some situations you can get it to work, but why bother?

is there sources for postgreSQL for dummies?

or nice comprehensive manuals?

or 3rd party books?

I have this book and it's excellently written. It's targeted at the beginner to intermediate level, and serves double purpose as a reasonably handy reference. It didn't cover some of the more advanced features in the depth I would have like.

A quick search of amazon for keyword postgresql turns up 9 other titles.

I dont know about MySQL/SAP/Interbase, but I do know that PostgreSQL v7.1.x *does* have SSL support if compiled with it. So that item is incorrect in the table.

Also, the table lists that postgresql has replication capabilities via erserver. This is only sort of true. Erserver is an asynchronous replication system, which is probably less than ideal. There are several replication projects going on for PgSQL right now (see here), but its most likely that WAL logs will somehow be used in the long run to provide real replication capabilities to PgSQL.

I would recommend trying PostgreSQL again. Relatively recently (v7.0 and v7.1) there were some major improvements in terms of speed at least. As to setting up PostgreSQL (i.e. create and configure a database), it is very easy, once used to it, and the documentation is pretty good (http://www.postgresql.org/idocs/ ).

I enjoy seeing MySQL becoming a mature database, but I do not exactly know what this hype is all about!? PostgreSQL supports server-side programming for a long time now (the same way it supported transactions long before!). At the moment it supports three different (and working) implementations:

- PL/pgSQL (this is the counterpart to ORACLE PL/SQL)
- PL/Tcl (TCL Procedural Language)
- PL/Perl (yes, this is server-side perl implementation!)

Here the pointer to the corresponding manual entry

Why would one use MySQL, while a powerful open-source DBMS has been available for years, and features stored procedures in any language (thru shared objects) ? AFAIK, it has always beat MySQL in benchmarks, too.

Want a true solution for an OO-capable DBMS ? http://www.postgresql.org/ :)

deepmind

Postgres home page

• New logo (check)
• Stored procedures in Perl (check)
• Stored procedures in C, C++, Python, TCL
• Langauge similar to PL/pgSQL
• User-defined datatypes
• Transactions
• Subqueries
• Constraints
• Stop being a bunch of whining Euro sue-boys

Looks like it might be a while. Better just get PostgreSQL in the meantime.

Well if he needs those two things, why not mention PostgreSQL to him? It can do both things - its a more feature complete DB than MySQL though in SOME cases its slower. But it all depends on teh application and just like WIndows vs Linux you'll see lots of MySQL vs Postgre flame wars too. I use both. I use MySQL for web based apps that have fairly simple backend needs and PostgreSQL for more complex setups.

Don't forget about PL/pgSQL, the procedural language for Postgres. This was written with Oracle converts in mind ;)

You mean that open source database MySQL pukes everytime it gets a little traffic?

Yes. And every once in a while, Slashdot and Everything really feel it. On the other hand, PostgreSQL (another free DBMS and the basis for Red Hat Database) supports proper ACID transactions and nested select statements and is known to lag quite a bit less in high-load situations.

<OT>
Also, the first seven initial caps on its web site spell "Pelt JTT" because he's a wooden actor.
</OT>

Have any OSS databases been ported to MacOSX?

• MySQL
• PostgreSQL

Ohhh, an OODB. Sweet, are there any OS ones yet?

According to postgresql.org, "PostgreSQL is a sophisticated Object-Relational DBMS, supporting almost all SQL constructs, including subselects, transactions, and user-defined types and functions. It is the most advanced open-source database available anywhere. Commercial Support is also available."

I was unable to find any software to create, review, store & update scientific knowledge on any particular subject. In hot areas of molecular biology new papers are published daily. No one has time to search and read everything that is relevant to his field of interest. It takes plenty of time to find most important papers, to obtain copies of it (in print or pdf) and to dig throw all unimportant parts of the paper that has to be included because you can not link (hyperlink) directly to certain paragraph in another paper.

I think that we need to switch from paper type publications (where earch paper consist of redundant data and is linked to other papers by references only) to knowledge-base like publishing. The knowledge base on any field of science may be stored in object oriented database with schema that is like table of contents taken from any "scientific bible" (i.e. "Genes VII"). The table of contents might be followed by another layer containing very condensed information (like in scientific reviews) that has links to titles of scientific publications, or better to one sentence summary of scientific finding of any particular publication. The next layer may consist of data taken from publication abstracts (containing everything impoprtant) with links to crude data. Each piece of crude data should be logically presented with rationale (reason for that experiment), discussion of crude result and links to exact paragraphs of other publications. Each fact stored in the database should be linked to crude data. The crude data should be taken from publications or direct submissions and of course before inclusion into the database it must be peer reviewed. The crude data may also include the probability of it's reliability (set by author or reviewers). The reliability is important because some results are obtained with methods that give reliable results, but many of the research in molecular biology provide only indirect evidence.

This scientific-knowledge database will remove most of the redundancy found in all scientific publications published as articles.

To create such a system you can start with object oriented database (i.e. ZOPE or object-relational PostgreSQL or MySQL that has good full text search). Please not that the database will contain trilions of hyperlinks, and link management will be very important part of it. It should have system for peer review (maybe similar to slashdot moderation). It will store crude results that are in hundreds of different formats. Each single crude result with additional data and files may be stored in one tar.gz or zip file. There must be also a some kind of versioning system. You will have to provide external editor or list of compatybile editors that suport XML format. Maybe OpenOffice? And last but not least - there must be a system that enable researcher to claim that this work is theirs. To obtain further funding scientist need to be able to show what research was done. In science the position of each researcher is based on how many papers he/her published in best scientific journals. Scientific journals has impact factro - if papers published in the journal are cited often - the impact factor of the journal is high.

-Piotr

PostgreSQL has some really nice support for proximity searching.

> second only in difficulty to programming enterprise-scale RDBMS systems.

You mean like Mysql?

Or perhaps PostgreSQL? NASA recently gave up Oracle in favor of MySQL, if I recall correctly. Besides which, on the gaming end, I recently purchased Tribes2, but I still find myself playing GLtron and Freeciv and Nethack more than anything else.

Surprisingly enough, all those games are also under the GPL.

The difficulty of making a game engine is exactly why Free Software is an ideal solution.

It remains to be seen how much further Oracle will go, and how quickly. However (and I've already prepared my will in case of God's heavenly thunderbolt), I predict that the Oracle database server itself will be open source within ten years. (If it isn't, it will no longer exist!)

As it stands, the postgresql project has been advancing very rapidly, and is getting extrememly close to beating Oracle in functionality. If Oracle doesn't react to this in the next few years, they're going to start taking serious hits in revenue.

Unfortunately, going open source is only *one* way they can react. Rather than going "open" they might go "free", that is, gratis. Using the db binary as a loss leader for other products (including support) wouldn't be the siliest thing they could do, and it would help to stay ahead of the *really* free competition for some time.

The Great Bridge benchmarks, Postgresql beat Oracle on speed slightly (Bruce Momjian confirms this in this interview: http://lwn.net/2001/features/Momjian/).

Version 7.1 has just been released, with the big news being the addition of outer-joins. I understand that replication is one of the next targets (and given the speed that these guys have been working at, I would expect it to be in the next rev).

From The postgresql site (http://www.postgresql.org):

Key New Features and Capabilities of Version 7.1 Include:

• Write-ahead Log (WAL) increases data integrity and processing speed. To maintain database consistency in case of an operating system crash, previous releases of PostgreSQL have forced all all data modifications to disk before each transaction commit. With WAL, only one log file must be flushed to disk, greatly improving performance. (Tech note: can eliminate use of -F in to disable disk flushes)
• TOAST (The Oversized-Attribute Storage Technique) Past releases had compiled-in row length limit typically between 8Kb & 32Kb. This restriction made storage of long text fields difficult, cumbersome and slow. TOAST enables rows of any length while maintaing the high performance PostgreSQL users have come to expect.
• SQL92 Outer Joins are now supported. (Tech note: eliminates the UNION/NOT IN workaround)
• 64-bit C Language Function Manager support The previous C function manager did not handle support 64-bit CPU's (e.g. Alpha, Sun, Itanium). (Tech note: This change should not impact existing custom functions developed for past versions, but performance will be improved through rewriting to use the new call interface.)
• Complex Queries that are better, stronger and faster Many complex queries were unsupported in previous releases. With v7.1 combinations of views, aggregates, UNION, LIMIT, cursors, subqueries, and inherited tables are enabled. Inherited tables are now accessed by default, and subqueries in FROM are now supported.

Pardon me for spoiling your FUD, but:

• PostgreSQL does supports these features.
• Interbase supports them too.

Maybe you should try using these databases before discounting them. Just because MySQL does not have transactions or row-level locking does not mean other 'free' RDBMS do not. As far as your claim that 'you tried'. What, exactly, did you try? Maybe the 'experts' you hired were stupid or incompetent. Maybe your architecture is inefficient. Maybe a hundred other things caused these problems and you incorrectly thought it was the fault of the database.
--
"In the land of the brave and the free, we defend our freedom with the GNU GPL."

There is some documentation work in progress on the topic here -> http://techdocs.postgresql.org/installguides.html# replicating

I've actually wanted one of those boxes for a while.

I would say that there is a LOT of potential there. You could have it synch with your addressbook and read the caller ID data. Thus, you would be able to tag who called and e-mail you that information.

I'd say that your best bet is to encode it in some sort of well-compressed file format, once you get it off of the modem. MPEG layer 3 would work just about anywhere, so you'd probably want to just use that as the file format.

I would say that for everything but the actual MP3 audio, and perhaps even the audio, you might just want to use PostGreSQL to store the data. It's accessible under Linux, and it's also accessible via ODBC with PsqlODBC. That way, you have the option for multiple interfaces. You can write a windows and Linux binary client, plus a web-client (Which is nice if you are at the office and want to check the home phone messages).

And, of course, check out Freshmeat to see if there's anything useful. I found KPhoneCenter and the VoiceModem Kit.

From PostgreSQL To Do List:

Under the title "EXOTIC FEATURES":
Add the concept of dataspaces/tablespaces

...it is not planned for 7.1

I'm just fine with PostgreSQL, thank you.

postgres works fine for me.

I would *never* think about being locked into a proprietary solution; Do you really take me as an idiot?

1. Oracle will support you. No matter what time of day it is, Oracle tech support mans the phones and consistantly is there to make things work again. That's of primary import for a company with a nearly unlimited checkbook.

2. Oracle is currently a little slower, but has a significantly larger feature set (most importantly, multiple versioning and transactions).

3. Wall Street like to hear you say "our Oracle order database allows us to blah blah blah..." Don't underestimate the power of the banker! There's a really good comparison at this site

Well, I don't know about mySQL, but GreatBridge provides 24x7x365 support for PostgreSQL, which is also open source and said to be more scalable than mySQL.

You can get an ODBC driver for PostgreSQL. I can attest that it worked fine on NT 4 when I worked with it. It is here. It's not too difficult; you just install it, then follow the instructions for configuring ODBC on the Windows machine, and create a database account on the Unix machine, which the windows machine can connect to, and you're basically done. It all worked fine for me at work this summer.

A database backend is incredible overkill for something like this.

I'd generally agree with you on this, but then I look at my own system and disagree.

I run PostgreSQL which is neither here nor there in this discussion, but the point I am trying to get back to making is that yes it's silly to use a fancy DB system for something like this but when you then also realize that I use the same DB to store more than that. I store info from my knowledgebase, vehicle maintenance log, family journals, MP3 database and a variety of other small projects within various tables on the same database system. The advantage starts to become clear when you realize that a single perl module (DBI::Pg) gets me access to any of this data and it's all (hopefully) stored in an efficent manner. I don't have to worry about parsing the text files differently or doing anything special to back them up. When you start to use the same DB system for more than one project the benefits of storing even silly little things in a DB start to show.

It's late, but I think it's kind of like realizing that 16k cluster sizes on the old FAT16 systems becomes much less hassle than when you partition and format to keep nice 4k or smaller clusters but then need 400M of room only to find that you ahve 50 here, 100 there, 75 somewhere else, etc... I long ago became bored of that and now use a nice LVM. I don't ever have to worry about running out of room or having to make a /usr1, /usr2, /usr3, etc. Same kind of thing with the database way of storing things that really don't need it.

You should do this kind of testing either to contribute to the kernel's development (this testing is more thorough than just casually trying it out), and especially if you're considering using a new kernel in a production system.

I welcome submissions of test suites to include in the article (or other articles to post at the Linux Quality Database - you'll see if you check out the site that there's not much there yet, but I have great hopes for doing good with it).

There have been some new suites submitted, including PostgreSQL's regression tests and one or two others that I have not yet added. I'll be updating the article soon.

Michael D. Crawford
GoingWare Inc

ftp://ftp.postgresql.org/pub/latest/

use LaTeX? want an online reference manager that

Don't you just love the way MySQL gives reasons not to use SQL features they don't support? And not just foreign keys, either, but transactions as well, among other things. Yeah, yeah, yeah, there's now rudimentary transaction support via BDB. Not that I'd trust any important data to something that is considered "rudimentary".

Then again, some people just have to reinvent the wheel several times over. One could wish that MySQL would just go away, in favor of more mature products that are also "Free". It won't happen anytime soon, but we can dream.

If find the big support for MySQL somewhat surprising. When I'm running a RDBMS, I want stability, even if it happen to be a commercial or a open source one.

MySQL have for many years been been considered as a very unstable product and that it have been (still are probably) lacking a lot of really useful features.

In the open source world, I find two RDBMS especially good. PostgreSQL and Interbase. They are both fine products and probably near commercial grade RDBMS standard.

So even if MySQL is now considered stable. So what? We have other choices, which have been developed with great care for many years and have been added features from their stability, not the opposite.

So in my humble opinion, if you want to run anything (semi)serious, definately do not run MySQL, there are MUCH better options out there.

I wrote a PostgreSQL book at http://www.postgresql.org/docs/awbook.html. The book is online, but I sold 2,155 print copies in the first month. Clearly having it online has helped book sales.

Did you consider PostgreSQL?. DB2's a fine database, I'm sure, but from your post it doesn't seem you are aware of alternatives to MySQL in the Open Source world... PostgreSQL 7 is a lot more fully featured than MySQL, and 7 is much faster than previous releases... faster than MySQL if certain benchmarks are to be believed. (I do not know, however I could understand if it scales better than MySQL...) For a little more info: Why OpenACS does not use MySQL.

Note: above post is not meant as a flame towards MySQL, I am just trying to point out possible alternatives which may have been overlooked. Thank you.

I'm surprised you are having so much trouble with Mozilla on a 650MHz Athlon. I run it on a 466MHz Celeron, 64MB RAM, and although it takes a bit more time than Netscape to load, its not unusuable.

I forgot up some href's ... so here they are:

Horde
IMP
IMAP
PostgreSQL