Upgrade from 6.9.2 to 6.10 - lost 1 disk in array, cache drive went offline

Format is NEVER part of rebuild. Format is a write operation. When you format a disk in the parity array, that write operation is treated just like any other, parity is updated. So after formatting a disk in the parity array, the only thing that can result from rebuilding that disk is a formatted disk.

 

Your diagnostics indicated filesystem corruption on disk2, cache, and the unassigned disk. None of these disks were missing or offline in your diagnostics, just corrupt.

 

Seems extremely unlikely upgrade related.

 

You should have asked for help in General Support before doing anything. Rebuilding was entirely the wrong approach to fixing your issues, and formatting any drive with data you want to keep is always a mistake.

 

There's constant crashing on the log, including references to the Unraid driver, there's another user that recently had similar issues with the same HP Microserver Gen8, including some data loss due to fs corruption with both btrfs and xfs devices:

 

 

For now I would recommend going back to v6.9.2, looks like these serves don't get along with the newer Kernel, unclear for now if all Microserver gen8 will have the same issues or just a specific CPU/BIOS combo.

Changed Status to Open

45 minutes ago, trurl said:

Format is NEVER part of rebuild. Format is a write operation. When you format a disk in the parity array, that write operation is treated just like any other, parity is updated. So after formatting a disk in the parity array, the only thing that can result from rebuilding that disk is a formatted disk.

 

Your diagnostics indicated filesystem corruption on disk2, cache, and the unassigned disk. None of these disks were missing or offline in your diagnostics, just corrupt.

 

Seems extremely unlikely upgrade related.

 

You should have asked for help in General Support before doing anything. Rebuilding was entirely the wrong approach to fixing your issues, and formatting any drive with data you want to keep is always a mistake.

 

So does this mean that I will indeed have to restore the array from a backup?

1 hour ago, Monteroman said:

So does this mean that I will indeed have to restore the array from a backup?

If you formatted disk2 and cache that is gone, but each of the other disks still has its data.

So I'll have to restore whatever was on disk2 then.  Cache, I knew I'd have to restore from backup. I've gone through that before. Well, it's going to be a long restore from crashplan then.. fun times ahead.

 

I've reverted back to 6.9.2. It's been stable on the MicroServer Gen8.