Bigbro's foray into the scary world of blogging
I finally managed to find the time to investigate the fate of neutron's disk array, the failing of which caused the whole machine to die unceremoniously a few days back. Turns out that it was a simple disk failure, which was simple to isolate and fix due to the disks being mounted in removable caddies. With my data stored as RAID5, the single disk failure was not a problem...
However: neutron is a cheap, IDE based system and this failure highlighted the differences between it and a commercial, carrier grade server. Although I lost no data, neutron would not successfully boot with the failed disk on the IDE bus.
The rebuild from degraded state of the large RAID5 array, upon which my data sits is almost completed now:
eason@neutron:~$ cat /proc/mdstat
Personalities : [raid1] [raid5]
md1 : active raid5 hda2[5] hde2[0] hdg2[4] hdb2[3] hdf2[1]
976623104 blocks level 5, 64k chunk, algorithm 2 [5/4] [UU_UU]
[================>....] recovery = 81.0% (197838464/244155776) finish=34.5min speed=22349K/sec
