Should WP drop MySQL 3.23 support?

WordPress currently requires MySQL 3.23.23 or greater. We’ve long avoid bumping our requirements so that we could accommodate the widest variety of hosting environments. We understand that people don’t want to upgrade a platform that is working just fine for them. However, supporting these older versions adds development and support costs. Recently, a commit went into trunk that bumped the MySQL requirement from 3.23.23 to 4.0 or greater. Considering that MySQL AB is dropping support for 3.23 and no longer provides binary downloads for 3.23, is it time for WordPress to move on? How many of you are still running 3.23?

Please note that this requirement change affects the upcoming 2.1 release only. The WordPress 2.0 series still supports 3.23 and will for the foreseeable future.

36 thoughts on “Should WP drop MySQL 3.23 support?

  1. If MySQL AB are dropping support for it, then it would seem more than reasonable that developers start to phase out support for it as well.

    Like you said, if you want/need/require a version of WordPress that supports a legacy database; then feel free to download an older, non-bleeding edgy copy of it. If that is the case, then maybe there should be an archved downloads page as well; listing all previous versions of WordPress and their basic requirements.


  2. It’s silly to support any product on a discontinued platform. When support for Win98 got dropped, vendors stopped making Win98 software. When MSSQL 7.0 got end-of-lifed, nobody developed against it anymore. WordPress should do exactly what other software vendors do and not support software that isn’t even available anymore. Asking if anyone is still using MySQL 3.23 is like asking who’s still using a 2.4 Linux kernel. There are a few holdouts, but it’s time for them to get with it and move forward or deal with not having the latest build.

  3. Why should support of an almost redundant program be continued, if MySQL AB are dropping their support for the product, why should WordPress still be developed to support it?
    Anyway, if MySQL AB isn’t going to support the product any more, then the webhosts should upgrade to continue receiving support. And therefore MySQL 3.23.23 support should stop. Would make the developers job much easier.

  4. Definitely drop MySQL 3.x support from the next version, IF it means that you’ll gain in performance (be it in development time, or in other places).

    After all, aren’t they on MySQL 5.x now?

    I’d think that for people on hosts using such an old version of MySQL, someone would bother writing in to ask them to at least move up to 4.x, after all.

  5. WP should have dropped MySQL 3.x support a long time ago.

    Unfortunately there is at least one popular web host (MediaTemple) which has a large number of customers on MySQL 3.x. (Their undeserved popularity, and numerous other problems with their service, are another topic entirely.)

  6. Most of the hosts I come across these days are running MySQL 4.x as it is. 4.0 more than 4.1, but that’s still very respectable.

    Supporting a platform that the maker of said platform no longer supports just doesn’t make sense to me. It’s time to move on.

  7. I am still running on a MySQL 3.23 version, but I have no problems if newer versions of WP would require a newer release. For me it’s a chicken and egg situation. I am running 3.23 because no application I use needs features from a newer version. If however WP 2.1 only runs 4.0 or higher, I will happily upgrade to that new platform.

    So, although it involves work for me, I have no objections against dropping support for the MySQL 3.xx software.

  8. The only problem I see with “there are older WP builds available for those running MySQL 3.X” is that upgrading WP is easy but downgrading is almost impossible. So, anyone on a MySQL 3.X platform trying an upgrade to WP 2.1 will probably just lose their blog (yes, “backup before upgrading”, but who does this ?)

  9. (ho and btw Ryan there is a problem with the loop that shows recent articles on your blog, shows same excerpt for all posts)

  10. Under normal circumstances I’d be totally in favour of dropping MySQL 3.23 support but..

    Many blog writers have their space hosted by a commercial outfit. It is impossible for these users to upgrade MySQL themselves and therefore to benefit from any improvements from dropping 3.23 support they will have to migrate.

    As this is the case surely it is worth looking at the absolute benefits WP would gain from movig to MySQL 4+ only compared to the number of uses who would either remain on WP 2.0.x or be forced to migrate to stay up to date.

    My host is on 3.23.58 – I’ve emailed them to ask about any proposed upgrades but I guess I’ll be sticking with 2.0.x or moving host when 2.1 arrives 🙁

  11. Doesn’t the benefits of MySQL 4.x (and 5.x) already makes the point moot?

    If somebody out there can enlighten me on the virtues of MySQL 3.x over the newer version, I’m definitely all ears.

  12. If MySQL is dropping support for 3.23, it’s likely because it’s just not practical to offer support across three full versions of their software. If it’s not practical for them to support their own software, why should other people?

    Perhaps when this happens, users will start complaining to MT (and other hosts running outdated software) and it’ll light a fire under their collective asses. It’s like Lammert’s chicken or egg situation, only by force.

    And of course there are always legacy versions.

  13. Only if you will REALLY gain some performance from this move, then do it in a next major version, but giving some support for an older ones (2.0.x) for lets say one more security bug-fix release, otherwise those unable to upgrade will move on to another blogging platform.

    I understand those who are voting to drop 3.23 as it being dropped by MySQL, but for the arguments sake i believe that the more platforms you support – the more users will support you.

    Announcing this move will motivate some of the hosting providers to move on as well, as they are keen to keep their blogger users =)

    I would say – announce it loudly so that everyone would know, and then in about 6 month do this move … =O)

  14. That mySQL are themselves dropping support for 3.23 reflects their priorities, but the WordPress team might have other priorities, like trying to reach a broad audience.

    That said, dropping support in upcoming WP releases may not narrow the audience for WordPress significantly, because 1) mySQL 3.23 users are likely to be a dwindling minority 2) there is reason to think that mySQL 3.23 users aren’t likely to upgrade to the latest and greatest versions of software.

    It is true that there are likely a number of 3.23 users because that is what their hosting provider offers, but they are perfectly free to change hosting providers. I don’t know for the fact, but I’d hope that the DB backup plugin makes this relatively easy.

    If it were my call to make, I’d drop 3.23 support in 2.1, but continue backporting security fixes to 2.0x so that people who don’t want to move (yet) don’t get screwed.

  15. Well my server is running MySQL 4.something so I don’t have a personal stake in this. I think it would be a good idea to create a ‘legacy’ version of WordPress (say the current release version for example) which receives ONLY security updates but can be used with MySQL 3.23. That way those who can’t upgrade their servers (for whatever reason) can continue to use WordPress (ie without having to expose their site to security flaws fixed in later releases) but the WP developers don’t have to spend time and money trying to get new features to work with an old database.

    Eventually such a version would become ‘security flaw free’ (ie there would be no new security flaws, due to no new features, so once any existing or undiscovered ones are fixed, that would be that) and therefore the cost (in time and money) to the WP devs of such a version would be nil after say a month or two of hunting down security flaws.

  16. As an update to my earlier post my host has announced they are moving their MySQL server solution from 3.23.58 to 5.0.26a.

    That’s a leap forward – so from my point of view drop support for 3.x if improvements are to be had.

  17. Well my personal perspective is that I don’t work with any web hosts using MySQL 3.23.x, so it won’t hurt me, but here’s my suggestion for a litmus test:

    MySQL is dropping all support for 3.23 as of the end of the year.

    If WP 2.1 comes out before the end of the year, (What are the chances!?!?) it would seem appropriate to support 3.23.

    For any version of WP coming out well after (~6 months?) support is dropped, even testing for 3.23 support is silly.

    If 2.1 comes out in early 2007, there is a judgement call to be made. (I do not see a target date for 2.1 on the page)

    And as for the question about improvements by dropping support–Another reason could be earlier releases because less stuff has to be dealt with/tested.

  18. If MySQL AB are dropping support for it, then it would seem more than reasonable that developers start to phase out support for it as well.

    That pretty much says it right there. I’m guessing that 3.23 comprises less than 2% of WP users, and most of those are likely people running their own server who will take this as an excuse to finally upgrade.

  19. Drop 3.x support.

    But a much more elegant solution would be to move all SQL operations from the different sourcecode files to one file.
    Thus creating a dedicated database layer. This layer could be easily adapted to any database out there.
    So the 3.x support could stay and new database engines could be supported.

    I am pretty sure someone had the same idea before and there might be some strong arguments against, but I hope that’s not true 😉

  20. If there is a benefit to dropping support for MySQL 3.23, then drop it. Benefits might include performance improvements, simpler SQL, features improvements, etc.

    But, please be sure that the upgrade script for WordPress checks the version of MySQL, and verifies that it’s current, and blocks the upgrade if it isn’t.

    There are lots of people who manage to install WordPress, but who would have no idea what version of MySQL they are running. You need to protect them from upgrading themselves into a problem.

  21. I’m running RH’s Fedora Core 3 still:
    $ rpm -q mysql

    We are planning to upgrade to FC6 shortly which has mysql.i386 5.0.27-1.fc6.

    However, upgrades are slow and painful. They take a lot of work and consideration. We don’t just want to drop in a custom mysql RPM, as we need security updates, but we don’t manage that ourselves. That means we take a huge step forward when we move to FC6 and not just mysql is going to be affected.

    FC3 is still supported by until FC7-Test1 ships (another month or so at least). Even Fedora has release maintenance/support cycles, and I think it’s totally acceptable for WP to do the same. Just give us all enough warning, and keep maintaining the WP2.0.x cycle for security patches (no new features) once WP2.1 ships to give those some more of us time.

    Webpages and blogging is a hobby for me, and not something I have a lot of time to devote on the backend to.

  22. Yups. Drop It!. Once its dropped, those never bothering upgrading it will ‘wake up’ a bit with ‘yawn’ n upgrade it.

    no harms.

  23. As long as there’s a comprehensive upgrade path, for WordPress, from MySQL 3.x to 4.x, it’s a good choice to move up to 4.x, especially if there are performance advantages.

    Bloggers who happen to use MySQL 3.x as their backend database can continue doing so for as long as they want… until they want to use the latest and greatest feature that’ll be developed for use with MySQL 4.x.

  24. Drop it!

    WordPress has always maintained security updates of previous versions, so users would in no way have a compromised/unsecure version. If you want the latest release though, you’ll need 4.x or higher.

    Keep moving forward…

  25. Drop 3.x support, the moreso as the 2.0.x series will keep running on 3.x.

    Most hosters offer the choice between 4.x and 5.x nowadays, I don’t know anyone still using MySQL 3.

  26. wp 2.1 doesn’t currently work with mysql 4 either, it uses mysql 5 only “match” “against” syntax. it could easily be changed to not use mysql 5 only syntax and work with 4 as well though.

  27. I definitely agree with many of the other guys here – since it is redundant, drop support for it. It doesn’t make sense to continue. Since MySQL AB are dropping support for their product, I don’t think WordPress ought to take on the responsibility of support.

    Kudo’s to you guys for keeping support for so long – that’s truly commendable. But I think in this case, you should not worry to much. You should upgrade to receive proper support. This isn’t a windows XP to windows Vista issue… lol.

Leave a Reply