Happy Hard Drive Joy
md1 : active raid5 hdg1[3] hdh1[2] hdf1[0] 156301184 blocks level 5, 32k chunk, algorithm 0 [3/2] [U_U] [>....................] recovery = 0.6% (513408/78150592) finish=254.4min speed=5084K/sec
Many thanks to Cat5 who picked up and dropped off a new hard drive for me. Decided to go with the larger 120G drive so I have a bit of extra space for the future. So my raid is happily rebuilding away. Thanks dude, next time I'll get the dancing girls for you like I promised.
Good news is that I found my business checkbook which for the life of me I couldn't find. Ended up finding it exactly where I thought it would be, not sure how I missed it. The thing I still haven't found is my toolkit with the soldering iron in it, which I need to find so I can solder the hard drive fan on which Cat5 found the problem (pulled wire on the back of one side. Thanks again, more dancing girls coming your way.
Nice…
I just did a dist-upgrade on a server (non-RAID) and it totally died:
hda: dma_intr: status=0x51 { DriveReady SeekComplete Error }
hda: dma_intr: error=0x40 { UncorrectableError }, LBAsect=31453484,
sector=31453484
end_request: I/O error, dev hda, sector 31453484
Eventually the kernel remounted the partion as readonly…
I wonder how a software RAID1 would have handled this. Ever had it where a drive kinda-sorta dies? does it get taken offline properly?
Normally when it gets errors like this it does take it offline properly. However, in some cases like I had, it doesn’t see that the drive is in error and drive requests just pile up (causing high load) as the drive isn’t being accessed. Doesn’t seem to cause problems other than not being able to kill processes (waiting for the drive) or unmount (drive is in use). In my case I had to shut down and deal with it offline. I’ve had the errors you describe on ufies more than once and software raid handles them just fine though.