Had a dual drive hiccup on my unraid server. My Parity 1 and Disk 19 both hiccupped at the same time. I shut the array down, attempted to restart to no avail.

I am running dual parity discs. I instructed the array to rebuild the first parity disc but I do not have emulated contents for the data disk, what is going on? Shouldn't my drive have been protected with dual parity even with 2 drives going down?

I have not touched the data drive that is being listed as unmountable. I am looking for advice on how to proceed while the Parity 1 disc rebuilds.

Disk 19 is not available in any way. I am getting a ZFS error on disc 19 on array start in the log. 

I have no idea how to proceed after the parity rebuild.

8 hours ago, Spazilton said:

but I do not have emulated contents for the data disk,

You have, it's just not mounting, Please post the diagnostics.

Thank you in advance for any assistance you can provide on this issue.

 

unraidserver-diagnostics-20211017-2146.zip

Check filesystem on disk19.

4 minutes ago, JorgeB said:

Check filesystem on disk19.

I'm assuming I should wait until the parity rebuild finishes? Since it needs to be in Maintenance mode to do the filesystem checks.

Edited October 20, 20214 yr by Spazilton

Yep.

Did the first step and it appears to have been successful to a point.

 

However when I attempted to start stray, brought it back down in maintenance mode and reran the check it appears to be bad.

 

After about 2000 lines 

 

it finished with

 

would have junked entry "PO'ed" in directory inode 6784074706 entry "Syphon Filter 3" in shortform directory 6784074711 references non-existent inode 351228389 would have junked entry "Syphon Filter 3" in directory inode 6784074711 entry "ThousandArms" in shortform directory 6784074712 references non-existent inode 351228390 would have junked entry "ThousandArms" in directory inode 6784074712 entry "TOMBRAIDER CHRONICLES" in shortform directory 6784074713 references non-existent inode 351228392 would have junked entry "TOMBRAIDER CHRONICLES" in directory inode 6784074713 entry "SLUS01011" in shortform directory 6784074728 references non-existent inode 351228410 would have junked entry "SLUS01011" in directory inode 6784074728 entry "SCUS94309" in shortform directory 6784074731 references non-existent inode 351228411 would have junked entry "SCUS94309" in directory inode 6784074731 entry "SLUS01019" in shortform directory 6784074747 references non-existent inode 351279749 would have junked entry "SLUS01019" in directory inode 6784074747 entry ".." at block 0 offset 80 in directory inode 6796019529 references non-existent inode 351279788 No modify flag set, skipping phase 5 Inode allocation btrees are too corrupted, skipping phases 6 and 7 No modify flag set, skipping filesystem flush and exiting.

unraidserver-diagnostics-20211021-0529.zip

 

I have another drive coming in this morning so that I can clone this drive before attempting any major repairs. I just want to preserve as much data as I possibly can.

 

xfs_repair with -nL flags shows the repair would not complete successfully if attempted.

 

Cost really isn’t object right now.  Just need to know how to proceed.

Edited October 21, 20214 yr by Spazilton

Run it again without -n, if it asks for it use -L.

On 10/21/2021 at 6:27 AM, JorgeB said:

Run it again without -n, if it asks for it use -L.

This did it. Problem solved. Tons of files put in to lost and found but it doesn't look like I lost too much. Thanks again for all the help JorgeB!!

Edited October 23, 20214 yr by Spazilton

If there's a lot on lost+found you can always mount the actual disk with UD (instead of the emulated one), and if that looks better you can rebuild parity instead of rebuilding the disk.

9 hours ago, JorgeB said:

If there's a lot on lost+found you can always mount the actual disk with UD (instead of the emulated one), and if that looks better you can rebuild parity instead of rebuilding the disk.

OMG! Thank you so much! The filesystem on disc when I unassigned and mounted it with the array down is perfect!! I'm fully backing up the contents to a spare drive and then will let it rebuild the drive.