I am running v6.3.3  A few months ago I noticed that I had disks with file system errors. I attempted to troubleshoot them and then life got in the way.  With a new baby now 2 weeks out, I’ve decided it’s time to get the server back online.  The problems is a bit overwhelming for me, there are multiple disks with file system errors.

 

I also remember, while running some of the disk maintenance before that the server would become unresponsive.  I plugged in a monitor directly to server and noticed an error on the console.  I can't find what that error was, and it hasn't happened in the past 2 days of re-attempting resurrection.

 

I have a mix of XFS and Reiserfs disks.  Tonight, I ran a disk check on all the drives, I’ve attached those results on the drives with errors (1,2,3,4,6) as well as a full diagnostics report.  Disk5 is the only one without errors.

 

If i had to devise a plan of action, it would likely the following:

Disk1, I’ve already run rebuild-tree 4 times and the results do not appear to be making much progress.  (I’ve attached all logs) / keep running rebuild-tree?

Disk6, run: xfs_repair -v

Disk4, locked up whole unraid during scan

• attempt another check.  Then likely --rebuild-tree

Disk3, run: reiserfsck --rebuild-tree

Disk2, run: reiserfsck --rebuild-tree

 

This forum has helped me once before with a disk issue, so I’m reaching out again before i do anything drastic (or incorrect).  

Unraid_08.21.2017.zip

It's not normal to get filesystem corruption on so many disks at the same time, run memtest for 24 hours before attempting any more repairs, or better yet, move the disks and the unRAID flash drive to a different pc and try to repair them there.

OK - I can try both of those things.

OK, so I thought I was onto something, based on the previous feedback.  That it was hardware related, not disk related, which I still think is true, but I don't understand what it could be.

 

I ran Memtest for a lot of hours, with no errors.

I took Disk1 out and loaded it into another PC, I was able to run rebuild-tree successfully.

 

This got me thinking that perhaps it is the SAT2-MV8 controller card in my SuperMicro server that is causing the problem.  In my server I have 8 HD bays going through 1-MV8 card, and 4 HD bays, connected direct to MB.  So I plugged in Disk1 direct to MB, no errors.  I plugged into a HD Bay that goes through MV8, it had errors.

I thought I found the problem and I have a spare MV8 card, so I swapped the cards.  Same results.

So I thought maybe the PCI Slot went bad, moved it to the other PCIX-133 slot on the board, same results.

I thought maybe the PCI-100 slots were on a different bus than the PCI-133 slots, so I swapped the card into a PCI-100 slot, same results.

Now, I thought i was going crazy so I connected Disk1 back direct to MB, and it shows no errors still.

 

I tested with other drives.  When connected through a MV8 card, they throw errors and recommend a rebuild-tree.  When I plug them directly to MB, no errors.

 

So besides jumping to the conclusion that my MotherBoard fried somehow.  Is there any other explanation?

 

FYI, here's my server specs:

• 12-bay Supermicro server
• Supermicro SAS826TQ
• Supermicro X7SBE
• Supermicro SAT2-MV8

 

any other ideas on what to look at next?

 

Thanks.

 

 

Disk1-Scans.zip

Probably a motherboard problem.

It's been a while, but I finally got a new motherboard, CPU, Memory, SATA Controller card installed.  I ran a Filesystem Check on all 6 disks and none came back with errors!  Woo Hoo!

I ran a Parity Check with the 'write corrections to disk' unchecked.  It returned 124,830 errors.  I'm guess this has to do with running the reiserdisk --rebuild-tree multiple times on disk1, like I mentioned in my first post.

 

It seems like I'm on the right track, but just wanted to check on the next best steps.  I'm guessing I should just write all corrections to parity, and I'll have to deal with whatever file loss the rebuild-tree process caused.

 

I've attached my Diagnostic File, I looked through this but don't see where it says what the 124k errors are....

 

If you've got thoughts on next steps, it'd be appreciated!

 

Thanks.

dwtower-diagnostics-20171028-1920.zip

5 hours ago, dewittig said:

I'm guess this has to do with running the reiserdisk --rebuild-tree multiple times on disk1, like I mentioned in my first post.

 

If the filesystem was checked like it's supposed to, by starting the array in maintenance mode and checking the mdX device, parity would be maintained, either way your only option now is to correct those sync errors.