With the release of CentOS 5.5 ext4 is considered stable in this distribution so I decided to migrate to it. Luckily I started from migrating fresh server with CentOS 5.5 using some instruction I found on the internet. I think I shouldn’t say, that I screwed the whole thing up 😉 After about 6 hours cursing, kicking, and crying I solved the task and figured the correct sequence of actions. The small problem with migrating root partition is that you can’t unmount it BTW.
May 07 2010
Recently I assembled a server with several 2Tb disks in Linux RAID. RAID resync reported by cat /proc/mdstat showed that resync will take about 9 hours! More than my working day! I started to look for a way to increase resync speed.
May 03 2010
Modifying my WP-SynHighlight addon for WordPress I came across a problem, that WordPress rewrites symbols inside my codesyntax shortcode with their nicer Unicode equivalents. I patched code against some. Then against some more. And then some more… And at last I felt tired of patching my code. Suddenly, I came across no_texturize_shortcodes filter, documentation for which at the time of written was published nowhere.
May 03 2010
Actually, I never bothered to optimize my server for VBulletin, because the server we purchased for forum is really powerful. But once upon a time logwatch told me, that smartd told her, that he found several relocated sectors on one of server HDDs. And then instantly I realized, that many time ago when I assembled that server, I configured HDDs to be in RAID-0 (striped mode)… Oh…. When you will make some stupid thing on your server, please keep in mind, that there are some people that can do things even more stupid…. Nothing critical happened, because I, actually, have all database backups. Database snapshots are made on server using this script of mine every hour or so (BTW, database backup on my server of 1.12Gb database (reported at admincp) takes around a minute to complete together with compression).
So, we had to visit datacenter and remove/replace HDDs. To allow our forum users to continue using forum until we fix that, we moved our forum to a new server. After forum move, I upgraded it to v4 to benefit from friendly URLs. And there it started. Server load almost immediately reached 4.0 (server had two dualcore CPUs, so this is almost 100% of standard load) and those were even not peak hours. Of course, that was not because of V4 (I turned off both that huge CMS issuing ~200 queries on home page and blogs). This was because that new server where the forum was moved to was less powerful. Of course, 4.0 is not a heavy load for server (once one of my servers reached around 100.0 and still I was able to login by SSH 🙂 ).
But nevertheless, that was the moment, where optimization hysteria first visited my head.