It’s been about a week since the initial launch of the first 5.0 packages and the OurDelta website, so it’s time for some first impressions and an update on what’s going on right now…
- Immediately lots of hits on the site (from 67 distinct countries), most downloads and repo retrieves started happening a few days later.
- Interest spiked even more when we put up the Debian 4.0 (Etch) packages, apparently there was serious pent-up demand there!
- Lots of positive responses from clients, user groups, global MySQL community, and in fact also from within Sun/MySQL. That’s very nice, as it provides confirmation we’re doing something worthwhile!
- No negative responses, blog posts or articles that I know of (and yes I did search around a bit through Google, Technorati, etc) which is of course nice too, but rather unusual and thus surprising. Isn’t there always someone who doesn’t agree? and anyway, a bit of dissent and discussion is healthy.
- We submitted sessions for OpenSQLCamp, the linux.conf.au Open Source Databases Miniconf, and the MySQL Conference and Expo 2009 call for papers. Not just about the project, but also about using the toolchain and delving into MySQL Server code for non-gurus.
- Noteworthy: MySQL Performance Blog is a popular site and so quite a few referrals came from Baron‘s post there.
- As expected some bug reports were received, mainly to do with packaging details. Steve Walsh, Peter Lieverdink and others have been very active sorting out these issues and they’ll be included in the next builds.
- No OurDelta-related external bug reports about server operation itself, while people have reported running the builds in production. Excellent!
- The ourdelta-developers mailing list got a fair number of new participants. Do join too!
- #ourdelta IRC channel (on Freenode) not massively busy but alive and well, today Peter Lieverdink added an IRC bot to notify of OurDelta bug updates from Launchpad there, which is very useful too.
- We received kind offers for mirroring the download/repo files, James Iseppi is working on the rsync and other details for that.
- While organising our infrastructure for better patch management (Quilt) and the build workflow, we did find some bugs in existing patches. Most have been fixed, some are still being discussed. It’s nothing major but, for instance, we don’t want patches to create any new reserved words as that can break existing applications.
- The above has slowed down our upcoming release of 5.1 somewhat. The intent was to make sure that all relevant 5.0 patches would also be available for the 5.1 builds, so that an upgrade would not lose you any features you’ve now come to like! Some porting work is involved there because not all are available for 5.1. With the other bugs getting fixed first, that’s eating up some time. We believe that’s ok though, as the resulting quality is higher.
- While working on one such issue (getting rid of startup warnings from one patch on 32-bit builds) we encountered an existing MySQL bug which had been hiding somewhat mislabeled. While the visible symptoms were benign, Antony Curtis was able to add insight to the root cause of things, which indicated that on some platform more serious problems could occur in the server. The story is essentially about old code and old platforms, but just in case, we made a patch and it’ll be in the next OurDelta build – to make sure it definitely can’t cause any trouble. All info and the patch have also been submitted to the MySQL bug, and the MySQL QA and bugs people have been very kind and helpful.
- Eric Bergen of Proven Scaling submitted a patch, and David Stainton from Spinn3r supplied two more that he’d extracted from the big Google patches.
- Other excited individuals have been delving into the repositories to play with various bits, or are doing some benchmarks. All good stuff!
Thus, you may agree, plenty going on so far and not at all bad for a first public week of a new project – and as is usually the case, a release doesn’t mean the work is done for a while, instead it immediately creates more!