Category ‘Internal’
A Friday 13th story
Are you superstitious? Do you believe in Murphy’s law? I usually don’t believe in such things, but sometimes you start to have doubts…
This site is served by my own virtual server hosted by Carrot Servers in Austria. The hardware nodes running their servers are housed in one of Germany’s largest hosting company’s datacenters. However, three of their datacenters recently suffered from a power loss because of a defunct uninterruptible power source and a series of other unfortunate problems – a little over 12,000 servers were without power for a few hours.
However, it seems my vServer’s hardware node was damaged by the power loss; one of the hard disk drives in the RAID array was replaced after the outage. Yesterday, the worst-case scenario came true: besides the hard disks, apparently the motherboard was also damaged causing the RAID controller to slowly corrupt the data on the drives. On Friday, 13th, my provider informed me data had been lost and the server will be swapped with a new machine.
Luckily, I did remember creating backup – and it happened to be a recent one (namely from Thursday, 12th). But as Murphy’s law states, things never work the way you expect them to work. In this case that meant that permissions were incorrect on the backup although I used rsync with the appropriate preserve-permissions option. I’ve thus been busy reinstalling the software, uploading the files and databases and fixing their permission. As of now, all data has been successfully restored (blessing in disguise) and I’m supervising the server’s functions.
Up or Out: employees in IT companies quitting
I just read an article in Alex’ Soapbox on IT professionals and their motives to quit a job at a certain company. Having just finished a 6-month-internship at a local new media agency this article came right on time. I don’t want to judge neither the article nor the company I’ve been working at, but the theory mentioned in “Up or Out: Solving the IT turnover crisis” is absolutely worth reading if you haven’t already.
Personally I think Alex Papadimoulis’ findings are a bit too black-and-white (as he admits near the end of the article: “Obviously, this article has painted some incredibly broad strokes, the largest being the stark dichotomy between skilled and unskilled developers, and the lack of distinction between organizations.”), but they are true nonetheless.
The smaller the error the bigger the bug
Bug tracking in software development can be a pain - especially when there doesn’t seem to be any logic behind the bug.
So we had that navigation tree at work consisting out of two category levels, for simplicity let’s call them Level-1-Category and Level-2-Category. Level-1-Categories could contain other Level-1-Categories and Level-2-Categories, while Level-2-Categories could only contain Level-2-Categories. The navigation was built, caching was implemented, everything seemed fine from a logic viewpoint, however it behaved weird where it would sometimes display the Level-2-Categories in a particular Level-1-Category and sometimes not or in different order.
When we finally found the bug (after hours of debugging) the scales fell off our eyes:
Now imagine what happened when you had a $level1Cat with ID 1 and a $level2Cat with ID 12, that would write it’s cache to
md5(‘112’); - and trying to open a $level1Cat with ID 11 and a $level2Cat with ID 2 it would read the cache md5(‘112’);. Unfortunately these classes did have the same interface, so no errors were thrown.Changing the code to
resolved the bug and finally fired the closing time-event for us.