A shabby week

· by Steve · Read in about 3 min · (529 Words)

The start of my week hasn’t gone that well at all. Firstly, I went to the dentist for a checkup which went rapidly downhill - unfortunately I had impacted wisdom teeth which were removed several years ago, but they left behind some hidden problems which were missed or deemed not serious enough until I had a routine x-ray on Monday, at which point my dentist says it now needs sorting out. I’ve had 2 appointments this week, and I have 2 (possibly 3) more later in the month to sort it all out. Given the problems are right at the back under the gums I’m not looking forward to those sessions I can tell you. There’s something very wrong about paying large sums of money to be put under extreme discomfort.

Secondly, I was trying to get the new ogre3d.org server configuration finished and I started hitting some weird problems. up2date randomly failing, Apache child processes falling over every 20 or so requests, and what I perceived to be serious underperformance when doing major work. The latter was linked to the server not picking up the second core on the box properly, although it seemed worse even than that. Some calls to support later and it eventually transpired that there was a BIOS issue with that motherboard / kernel combination that meant not only did it fail picking up the dual core chipset properly, it had ended up making things terminally unstable, and we couldn’t be sure the system libs hadn’t been screwed up during up2date failures. The safest option was to reinstall the entire box, sorting out the BIOS / kernel issues right at the beginning. It took them a while, but they sorted it all out - but of course this set me back several days since I had to repeat my setup tasks (although my provider did install all the extra base software this time).

Now, I’m not really annoyed with the host particularly since I’ve been in that position myself before as tech support, where you just get unlucky with a new piece of hardware, new BIOS issues that don’t manifest themselves immediately etc, and they have been genuinely helpful throughout - one engineer staying until late in the evening to get it all done. I’m also lucky this was discovered before the site was live, since the interruption would have been disastrous had we already moved. But nevertheless, this has cost me time, stress and inconvenience that would have been avoided had the original server as delivered been trouble-free, and managed hosting isn’t at all cheap (even with the discounts I managed to negotiate), so I’m claiming for service credit as compensation, which is covered under the SLA I have anyway. According to my reading of it I should get one or two months free.

All these things have meant little OGRE coding has got done so far this week so I apologise for things that are backing up. I’m hoping to get the new server finalised today, at which point I’ll run it in for a while before organising a switchover early next week. If all goes well, I think you’ll like it.