XFS Meta Data Corrupt

[gzibell]

Hey All, 

 

So first real issue with my unraid server happened twice yesterday.  Everything has been running perfectly for a couple of years now.  All services stopped, VM's, Dockers, everything.  Couldn't access webui or ping the server.  Connected a monitor to a it to see what was up and there was a batch load of numbers and the only real thing it said was, 

 

XFS (md2): metadata corruption detected at xfs_dinode_verify+......

XFS (md2): unmount and run xfs_repair

XFS (md2): First 128 bytes of corrupted metadata buffer:......

 

So I hit the power button, it went through clean shutdown, fired up again and back to normal.  Anyone know what this is, how to fix it, or where to start looking for answers?  Any help would be much appreciated.  Running 6.6.0

[Squid]

Diagnostics would be better, but based on the snip you need to Check Disk Filesystem on disk 2

[gzibell]

Any specific part of the diagnostics or the whole zip file?  

[itimpi]

6 minutes ago, gzibell said:

Any specific part of the diagnostics or the whole zip file?  

Normally the whole zip file.   It is never clear which bit is going to be the one that provides crucial diagnostic problem for a particular problem.

[gzibell]

Ok so I ran xfs_repair on disk 2 in maintenance mode on disk 2 first with -n and many errors were reported.  I then re-ran as -v and it looked as though all errors were corrected.  Re-running as -n did not report errors again. Rebooted array and all services came back online. 

 

So if that it?  Error corrected?  Do I need to check files on disk 2 or anything like that?  I guess I am just not sure what xfs_repair does. 

[gzibell]

Alright guys, one more issue.  after the repair all seemed to be normal. At some today the VM i have setup remotely manage my server via team viewer went offline.  So i connected to the vpn on my usg and was unable to ping, or ssh to my unraid server.  

 

Get home from work and check the monitor and all looks normal.  Sitting on the unraid login.  Didn't know exactly what to do so hit the power button, waited for clean shutdown and rebooted.  All back up and running again.  Checked disks with xfs_repair again and no errors reported.  

 

Maybe this was a fluke?  Have had no issues up until I upgraded to 6.6.0 and 6.6.0 ran fine for a few days at least without issue.   Might be nothing but trying to get to the bottom of it.