I was performing a parity check in 4.7 pro and I saw this in the syslog:

 

Jul 25 18:24:36 Tower kernel: handle_stripe write error: 22259944/2, count: 1

Jul 25 18:24:36 Tower kernel: md: disk2 write error

Jul 25 18:24:36 Tower kernel: handle_stripe write error: 22259952/2, count: 1

Jul 25 18:24:36 Tower kernel: md: disk2 write error

Jul 25 18:24:36 Tower kernel: handle_stripe write error: 22259960/2, count: 1

Jul 25 18:24:36 Tower kernel: md: disk2 write error

Jul 25 18:24:36 Tower kernel: handle_stripe write error: 22259968/2, count: 1

Jul 25 18:24:46 Tower kernel: hdb: status error: status=0x7f { DriveReady DeviceFault SeekComplete DataRequest CorrectedError Index Error }

Jul 25 18:24:46 Tower kernel: hdb: status error: error=0x7f { DriveStatusError UncorrectableError SectorIdNotFound TrackZeroNotFound AddrMarkNotFound }, LBAsect=140183580147464, sector=18446744073709551615

Jul 25 18:24:46 Tower kernel: hdb: possibly failed opcode: 0x50

Jul 25 18:24:46 Tower kernel: hdb: drive not ready for command

Jul 25 18:24:46 Tower kernel: hdb: status error: status=0x7f { DriveReady DeviceFault SeekComplete DataRequest CorrectedError Index Error }

Jul 25 18:24:46 Tower kernel: hdb: status error: error=0x7f { DriveStatusError UncorrectableError SectorIdNotFound TrackZeroNotFound AddrMarkNotFound }, LBAsect=140183580147464, sector=18446744073709551615

Jul 25 18:24:46 Tower kernel: hdb: possibly failed opcode: 0x50

Jul 25 18:24:46 Tower kernel: hdb: drive not ready for command

Jul 25 18:30:01 Tower kernel: hdb: irq timeout: status=0xd0 { Busy }

Jul 25 18:30:01 Tower kernel: hdb: possibly failed opcode: 0xb0

Jul 25 18:30:01 Tower kernel: hdb: status timeout: status=0xd0 { Busy }

Jul 25 18:30:01 Tower kernel: hdb: possibly failed opcode: 0xb0

Jul 25 18:30:01 Tower kernel: hdb: drive not ready for command

Jul 25 18:30:01 Tower emhttp: disk_temperature: ATTR_Temperature_Celsius not found

 

Unraid stopped doing anything and I clicked the "stop" button to take the array off-line.  Now I'm being told to wait and I can't connect to the Webgui and its not responsive to anything.  But I can still telnet into the box.  Not sure what next steps to take.

 

Ultimately, I have a spare drive that I can move files onto, but it's already in the array and formatted, so I don't think I can count on the parity drive to recover to it.  And I don't think I have another old IDE of the same size I can stick in there.  Any suggestions?

 

I've attached the output of my smartctl  -a  -d  ata  /dev/hdb command (rpt1) and smartctl  -a  -d  ata  /dev/hdb as rpt2.

 

I've followed the instructions on http://www.lime-technology.com/wiki/index.php?title=Check_Disk_Filesystems and reiserfsck returned no errors.

 

I couldn't powerdown cleanly so I hit the big red switch and upon reboot it immediately found 14 sync errors and is happily checking the remainder of all the disks.

 

So after reboot, unraid ran the parity check again.  And again there's some sort of error on my HDB disk.  I think I'll see if I can move stuff from HDB to one of the other drives and then eventually replace that drive.

 

unraid_smartctl-rpt1.txt

unraid_smartctl-rpt2.txt

Still struggling to recover data off HDB.  I can only interact with unraid through a telnet session. 

 

Right now the data pull off the failing drive is averaging about 2.2MB/s.  That's really slow and I'm starting to think maybe the parity drive is at risk? 

 

I did a top command and I see the top 2 activities are unraidd and mc.  I'm using mc to manually move files off HDB so that I can at least keep track of what it's doing.  I did try the the nohup mv /mnt/disk2 /mnt/disk7 & trick last night, but it stopped transferring files at around 2am last night.

 

Here's hoping this type of rescue operation works!

You can replace an IDE drive with a SATA dive if you have a spare SATA port. unRAID does not care about IDE vs SATA.

Power down and disconnect the bad drive. The parity will simulate the drive and without the bad drive causing issues any copying should go much quicker.

 

And yes, you could replace the IDE with a SATA drive.

 

Peter

Thanks Lionelhutz...

 

In the end I managed to get all the data off the drive.  I've now gone out and bought four 2TB WD20EARS drives, and I now have a long weekend to build a new parity drive and migrate data to the new drives.

 

2tb drives are awfully big aren't they? Takes 12 hours just to build a parity drive?!  Yikes.