Yesterday I replaced a drive with new one and started with a data rebuild on it. Woke this morning to another drive with a red x and 2 other drives showing errors but not red balled. 

What would the steps be to fix this mess? 

Replace the red ball drive and rebuild that drive and leave the other drives alone until after that rebuild? Not sure what route I should go.

Another question would be about parity checks.  Which is the preferred method, with correction or not?

tower-diagnostics-20180921-0856.zip

Edited September 22, 20187 yr by Harro
solved

Disks looks fine, at first there were read errors only on disk14, and second parity corrected those, but then there were also on disk13 at the same time, so rebuilt disk will have some corruption.

 

All 3 disks with errors are on the same miniSAS cable of an LSI HBA, so maybe a cable issue, possibly an HBA port issue, cable is more likely, it could also be power related, i.e., if those disks for example share a power cable or splitter.

 

IMO, best Way forward would be to check/replace/swap those cables and rebuild the disabled disk4 and rebuild disk1 again to fix the corruption on the first rebuild.

 

 

 

 

 

 

I started an extended smart report on 13 and 14 a few minutes ago. And yes the rebuilt disk1 has 4000+ errors on it.

Can I rebuild 2 disks at the same time, since I have dual parity? 

Just now, Harro said:

Can I rebuild 2 disks at the same time, since I have dual parity? 

Yes, just need to disable it by starting the array with disk1 unassigned, check that the emulated disk mounts correctly before rebuilding on top.

Rebuilding now and will report back in 17 hours.

Thanks johnnie

After 18 hours, rebuild complete without errors. I had replaced the miniSAS cable with a new one and rebuilt the 2 disks at once. Thank you again Johnnie.

👍