red balled disk

[downloadski]

One of my 4 TB hdds redballed in a parity check. (it was valid, but i wanted to see the parity performance)

 

I did change the tunable paramters to 4x the original value.

Attached the syslog, and did tests on the disc from the gui.

 

I am following the procedure Joe L. wrote down in a other thread about redballed drives

 

Stop array.

Unassign drive.

Start array with it unassigned.

Stop array once more.

Reassign drive.

Start array and let the drive get reconstructed.

 

 

It is now rebuilding.

 

I guess i should run parity check with nocorrect next time.

 

 

Can anyone tell me what the error here was ?

 

It looks to me the mvsas driver timed out:

Dec 28 14:33:59 Tower2 kernel: md: recovery thread woken up ...

Dec 28 14:33:59 Tower2 kernel: md: recovery thread checking parity...

Dec 28 14:33:59 Tower2 kernel: md: using 3840k window, over a total of 3907018532 blocks.

Dec 28 14:36:44 Tower2 fan_speed.sh: Highest disk drive temp is: 38C

[7m--More--(62%)[27m

Dec 28 14:36:44 Tower2 fan_speed.sh: Changing disk drive fan speed from: [153 (76% @ 909 rpm) ] to: [157 (78% @ 935 rpm) ]

Dec 28 14:43:05 Tower2 fan_speed.sh: Highest disk drive temp is: 39C

Dec 28 14:43:05 Tower2 fan_speed.sh: Changing disk drive fan speed from: [157 (78% @ 932 rpm) ] to: [161 (80% @ 967 rpm) ]

Dec 28 14:47:12 Tower2 kernel: drivers/scsi/mvsas/mv_94xx.c 626:command active FFDF7FFB,  slot [f].

Dec 28 14:47:12 Tower2 kernel: drivers/scsi/mvsas/mv_94xx.c 626:command active FDFFFFFE,  slot [39].

Dec 28 14:47:12 Tower2 kernel: sd 9:0:0:0: [sdn] command ec2a1180 timed out

Dec 28 14:47:12 Tower2 kernel: drivers/scsi/mvsas/mv_94xx.c 626:command active FFFFFBFF,  slot [4a].

Dec 28 14:47:12 Tower2 kernel: sd 9:0:0:0: [sdn] command de122300 timed out

Dec 28 14:47:12 Tower2 kernel: drivers/scsi/mvsas/mv_94xx.c 626:command active BFFFFBFF,  slot [5e].

Dec 28 14:47:12 Tower2 kernel: sd 9:0:0:0: [sdn] command deaa30c0 timed out

Dec 28 14:47:12 Tower2 kernel: drivers/scsi/mvsas/mv_94xx.c 626:command active 0FFFFFF2,  slot [62].

Dec 28 14:47:12 Tower2 kernel: sd 9:0:0:0: [sdn] command de9aa780 timed out

Dec 28 14:47:12 Tower2 kernel: drivers/scsi/mvsas/mv_94xx.c 626:command active FDFFFFFC,  slot [21].

Dec 28 14:47:12 Tower2 kernel: sd 9:0:0:0: [sdn] command dfa38c00 timed out

redballed.zip

[downloadski]

Rebuilt seems to go ok, now at 39%

I guess it might have been a slightly loose cable. I installed the power cable to my new power switch, so the case was open and on its side, and i wired the cable from the switch to the mainbord under the sata cables. I pressed on all the cables once more before i started the re-built.

[downloadski]

Parity check with not updating the parity disk results in over 9 million writes till now.

Parity check is at 19% currently.

 

I guess i need to download a lot that might be lost beacuase of this unraid behaviour.

Starting to dislike it again.

[dikkiedirk]

You did say you changed some values to 4x the original value. Maybe not such a wise decision. You can blame it on unraid, but it might be user error.

 

 

Also the SMART report you posted shows that that disk has reached a temp of 68 degrees. Not good. The disk might be toast (baked), no pun intended.

[downloadski]

Yes that 68 was because the 2 Fans for the hdd's not working for a few minutes.

[dikkiedirk]

Do you still trust that disk? I think I wouldn't. Maybe RMA?

[downloadski]

Do you still trust that disk? I think I wouldn't. Maybe RMA?

I do not trust my unraid server at all, lost enough data already..

[downloadski]

Parity check done,0 issues, so that rebuilt went ok.