3 disk errors RAIDZ2

[BonzTM]

Hey all,

I'm in a pickle and need some advice from the ZFS gurus.

I've got 2 10 drive RAIDZ2's in my pool. 3 disks from one of the vdevs are throwing me read/write and checksum errors. I've swapped cables and ports on the HP SAS Expander to no avail. The only thing other than drives I haven't ruled out is the backplane, and I don't have a spare to test.

I've got 3 drives coming in from RMA, but I'm not sure the best way to go about making sure I don't lose or corrupt the least amount of data.

Should I throw 1 or 2 of the drives in as a hotspare and let it rebuild when the drives fault out for error tolerance?

Need a little help. Thanks!

 

[mwroobel]

You do have up to date backups of all the data, correct? Can you please paste the results from the following:

zpool list
zpool status -x -v <poolname>
zpool history

 

[BonzTM]

Unfortunately I can't back up 20+TB of data. I have backups of about the most important 1.5TB of data.
I turned the server off until I get the disks in from RMA to prevent any problems. I will report back with the answers though.

 

[patrickdk]

Best bet it to throw a drive in, and to a zpool replace with one that is acting up, when that finishs, do another replace with another drive, till all bad drives are replaced.

The only way to loose data that way, is if 3 drives can't read the *same* location. So it will be unlikely you will loose data this way.