Hey all, I'm looking for the best solution..

 

I've been fighting a LOT of dead drives with my server.  I've since bought a new PSU but I'm still recovering.  Basically, I had two simultaneous drive failures at the same time, a 2TB and 500GB.  I manged to save the data by reading it through linux so that's good.  The data is stored on another disk but I'm waiting to get my array up 100% before I get the data back on it.  The 2TB drive failed first so I had that replaced and while that was going in the 500GB died.  I put in another 500GB but it had failed as well so my backup-backup plan was my 750GB.

 

I precleared the drive and the new 2TB and they all passed.  I than put them in my array, added the devices and went to format them as they showed up unformated.  This also failed; getting the bread: Cannot read the block (0): (Input/output error). on BOTH disks.  Attached is the syslog.  The first time I pre-cleared the drive was on my old PSU.  Pre-clear didn't seem to pick up the block (0) issue, I guess it just looks at the before/after conditions and if nothing has changed it believes it's okay. Is this how it works?

 

So, my array is running right now pretty much 'unprotected' since I can't do a rebuild if another disk fails so what's the best way out of this?  I'm going to RMA the disks as quickly as possible, and hopefully that fixes it but is there any other ideas?  Can I just delete the two disks (using initconfig), rebuild the parity based on the current drive and prey nothing dies on the rebuild and add them both later?  There is no data to be rebuilt on the two disks, I'd rather just add them in.

 

Thanks

syslog-2011-12-20.zip

Run initconfig to reset the array. Add the known good drives and rebuild parity. Then do a parity check. Run pre-clear on the two questionable drives before re-adding  them to the array.

pre-clear instructions