Links and notes
As if I needed another page for unsorted data.
- Hypertable
http://www.nabble.com/Slony-II-documentation-td15311818.html
http://kotaku.com/gaming/media-criticism/ken-kutaragi-misunderstood-or-just-nuts-111634.php
http://www.eurogamer.net/articles/hirai-were-the-official-industry-leader
http://blog.uncool.in/2009/01/19/computer-science-fail-higher-education-in-india/
http://peb.pl/showthread.php?s=1ef2cacc5b6f2cc1c8d2c3338874bfdd&p=1837090#post1837090
Fallout Gensokyo, some fantastic crossover writing - latest thread and storylist index with links to earlier instalments
Sake
05:30:13 < gozu> not_me: the supermicro cases suck really, really hard at isolating the disks from each other. 05:30:36 < gozu> basicly, there is a slim sheet of metal in between, and the bay-trays are not even tight fits. 05:30:50 < not_me> we have a few of those cases 05:31:12 < gozu> so we did detailed benchmarks on a disk, with the 9 disks "around" it being a) spun down, b) spinning but idle, c) active. 05:31:18 < gozu> and the results were superscary. 05:31:33 < gozu> more so when the disk has low cpu power. 05:31:53 < gozu> (== fails at head servo control) 05:32:06 < not_me> you mentioned that 05:32:15 < not_me> (the cpu power issue) 05:32:31 < kakazza> cpu power, hdds, what? 05:33:30 < not_me> kakazza: "server" hdds have more cpu power, to do more sensor data analysis to compensate for vibrations 05:33:55 < not_me> was one of gozu's findings when talking to techs of some hdd manufacturer 05:35:23 < kakazza> heh, nice. 05:37:07 < not_me> gozu: wow, they got a new model: http://supermicro.com/products/chassis/4U/847/SC847E26-RJBOD1.cfm 05:37:27 < gozu> they actualy seem to simply have two "cpus" on the "serious" hdds. one for talking to the disk, one for talking to the host.