Reports from the MySQL Users Conference

Eh-Wire writes "OnLamp is reporting on the MySQL Users Conference that is currently underway. Among the highlights are the announcement that the code for MySQL 5.0 is now complete. Axmark and Widenius suggest that squashing bugs is the key behind the success of MySQL. Michael Tiemann from Red Hat and the OSI delivered a keynote on "Defining Open Source". He suggests that Microsoft's "shared source license" has been a complete failure at the design level."

Comments

MySQL's New Slogan

[Anonymous Coward]

"We're catchin' up!"

...Reports from the MySQL Users Conference

[waldoiverson]

i don't think i've ever seen MySQL output a report like this...must have been a weird query.

Squashing Bugs Good?

[mwkaufman]

Who knew? Maybe Microsoft should follow this brilliant business plan.

Objection, redundant

[Tackhead]

> Microsoft's "shared source license" has been a complete failure at the design level.

Due to the rising cost of energy, ink, and/or toner, we urge all authors to reduce their word count wherever possible. For instance, the string ('s "shared source license") in the preceding article is redundant and may be eliminated.

Thank you for your co-operation.
- The Management

From the Article

[DarkHelmet]

"MySQL 5.0 is code complete"

So does this mean they're still doing bugfixes? Or they're doing testing? Or it's going to come out any moment?

Or does it mean that they're no longer implementing new features, which means they're in beta. If that's the case, we knew that much already.

Mysql needs to Improve

[RedElf]

It still has a long way to go, over the last weekend I spent several hours trying to recover data from a backup (mysqldump) because of dataloss due to mysql database corruption on server version 4.0.20.

Sure it seems ok if you want to run a small, fast, dynamic website that doesn't contain any real valueable data.

For serious data needs, in my projects I'll choose Sql-Server over mysql any day of every week. I'm not saying Sql-server is the best choice, but its ahead of mysql when it comes to data integrity.

Re:Mysql needs to Improve

[TexVex]

Data integrity is the database ensuring that when you want to have an entry in another table for each row in a table, that you do.

That's referential integrity. For example, if your database enforces referential integrity between an accounts payable and a payable entities table, if you delete a payable entity then all associated payables records would be cascade-deleted along with it.

And, as an aside, generally if you're going to maintain a record-for-record relationship between two tables, you might as well combine them into a single table.

Data integrity is making sure your data is complete and accurate. The database can assist with this by the enforcement of rules when entering, deleting, and updating information. Extending the above example, you might preserve data integrity by disallowing deletion of a record from the payable entities table when there is a non-zero balance on that account or there has been activity on that account in the past three years.

Re:Mysql needs to Improve

[Michalson]

Your perception of "data integrity" seems very limited, either that or you don't a thing and you're just bullshitting based on your knowledge of "data integrity" in other fields.

In the professional world of databases data integrity is ensured by the ACID compliance of the database. In your post though you only seem to understand the C in ACID, "Consistency". The problem is that out of the box MySQL will often fail the Atomicity and Durability requirements of ACID (and doesn't even attempt I). That means the database or a table can be left in a corrupt, unrecoverable state if interrupted by any number of things (you don't just lose the latest additions, you lose everything, forcing you to restore from a hard backup as grandparent did).

To properly understand this, look at another area where ACID is used. If you are using a disk formatted with FAT32 or ext2 and your computer crashes, you'll end up being forced to run scandisk/e2fsck, which has to go over the entire disk structure looking for errors. In many cases the errors it finds (due to an operation being interrupted leaving the disk in a corrupt state) are not recoverable and if left unchanged will mess things up even more. So you get little chunks of the disk (CHK files when using scandisk) that the system has no idea what to do with (databases are far more interconnected, making one unrecognizable element break the whole thing). On the other hand more modern disk systems (NTFS and ext3) borrow methods from the database world, resulting in a disk system that is ACID compliant. When the system crashes or is rebooted by a power failure, it doesn't spend 20 minutes checking the disk for errors it won't even be able properly fix, it just uses it's Durability feature to "instantly" restore the disk to the last good state (on a standard ext3 formatted 7200rpm drive "instantly" is less then 2 seconds).

The only way to make MySQL ACID compliant (and thus durable enough to store information that's actually important) is to use InnoDB tables instead of MySQL's fast default MyISAM tables. Even then, by InnoDB's own documentation, its particular form of ACID compliance is heavily dependent on how and when the OS writes to disk - you can still end up with unrecoverable InnoDB tables if the system crashes, making InnoDB more a device for its transactional programming capabilities and row level locking then it's long term data integrity features (A, C and I). With real world databases a crash simply means you lose the latest changes, not the whole table (to be redundent, the D in ACID, Durability).

People give MySQL too much crap...

[Anonymous Coward]

Really, it sort of annoys me. We use MySQL in a live, production environment. We have tables with close to five billion rows in them. MySQL has only given us a problem once. A table was mysteriously corrupted about two years ago. The fix? I opened up a HEX editor and repaired the damage. Restarted MySQL, ran some checks with the included tools, and we were all done. Quicker than pulling tapes, let me tell you.

I am really tired of all the hate around here. Every other comment is "If you're serious you will use Postgres because it has feature x, y, and z which make it a better product and you will suffer the consequences if you use MySQL because it's a bad thing for bad people because they don't like penguins and hot grits and bacon shit!"

Actually, it's not that bad... but it's damn near.

Re:People give MySQL too much crap...

[vadim_t]

Well, it might just be me, but I definitely don't look forward to hex editing databases when the thing goes down and the whole company grinds to a halt.

Now, I will readily say that mySQL has its uses. If you're simply aiming to logging data it works well. IIRC, that's what it was designed for in the beginning.

But it's still quite far from being a good database. Doing crap like silently ignoring things it doesn't like is one of the things that makes sure I'll never use it for anything important.

Re:People give MySQL too much crap...

[Anonymous Coward]

When the number of items on the "gotchas" list for MySQL (http://sql-info.de/mysql/gotchas.html) is less than or equal to the same list for PostgreSQL (http://sql-info.de/postgresql/postgres-gotchas.ht ml) I'll consider using MySQL.

currently underway, eh?

[donutz]

"OnLamp is reporting on the MySQL Users Conference that is currently underway."

But the conference website says it finished yesterday...