[SOLVED] Duplicate uuid problem

[superloopy1]

Guys, I have removed a drive from my main server having rebuilt to a larger disk. I now wanted to insert this drive into my backup server, complete with contents, as a "backup".  Done this many a time by just new config'ing after inserting said drive. This time I gat the dreaded 'unmountable no file system' so stopped array, removed drive, restarted, new configged again to get back to where I was and plugged drive in via UD. It mounted ok, xfs, contents visible. Unplugged for a bit of thinking time. Tried again and am now getting 'duplicate uuid' message. The drive only contains the base files which were rebuilt on my main server but I just need to understand why the procedure didn't work this time around. Would be nice to just slot it in again and see it's filestore as per the three other drives I've previously moved in this manner. Any ideas? Will regenerating it's uuid affect the disk contents? I'm assuming that the duplicate uuid issue is as a result of plugging/replugging the drive in at some stage which it didnt quite like so once i get over that then the issue of just why i got an 'unmountable fs system' and supposed drive needing a reformat (i didnt !) will still be there?

Edited February 1, 2020 by superloopy1

[superloopy1]

Here's diagnostics ....

xtower-diagnostics-20200201-1550.zip

[superloopy1]

Just tried a powerdown/restart and plugged drive in again via usb ...

 

 

'Feb 1 16:30:25 xtower kernel: sd 8:0:0:0: [sdl] Attached SCSI disk
Feb 1 16:30:36 xtower unassigned.devices: Adding disk '/dev/sdl1'...
Feb 1 16:30:36 xtower unassigned.devices: Mount drive command: /sbin/mount -t xfs -o rw,noatime,nodiratime '/dev/sdl1' '/mnt/disks/ST8000AS0002-1NA17Z_Z84127RD'
Feb 1 16:30:36 xtower kernel: XFS (sdl1): Filesystem has duplicate UUID 45f85bea-fde1-4ea2-bdf1-f5f56cf3f838 - can't mount
Feb 1 16:30:36 xtower unassigned.devices: Mount of '/dev/sdl1' failed. Error message: mount: /mnt/disks/ST8000AS0002-1NA17Z_Z84127RD: wrong fs type, bad option, bad superblock on /dev/sdl1, missing codepage or helper program, or other error.

 

So, seems that my current uuid is being bothersome ;>)

Edited February 1, 2020 by superloopy1

[itimpi]

You can get into this position if the disk has ever been rebuilt

 

If you go to Settings >> Unassigned Devices then there is an option there to change the disk UUID.

[superloopy1]

This disk hasnt been rebuilt though. I removed it from my first server and allowed the contents to be rebuilt onto a larger drive before attempting to install this original drive to my second server. Just whack it in, new-config, rebuild parity is how the previous three drives i've 'moved' have been done without any probs. This one came up 'no filesystem' for some reason. I backed out, took it out and then mounted via UD externally. This showed disk and fs was ok but subsequently came up with duplicate uuid. So, if i change uuid will i corrupt any data???

 

Still dont understand how uuid can be a duplicate, these are two separate systems its been mounted onto.

Edited February 1, 2020 by superloopy1

[JorgeB]

Likely another disk on the backup server came from the other one with the same filesystem, duplicate UUID can't be a coincidence, just generate a new one and you're done.

[superloopy1]

3 hours ago, johnnie.black said:

Likely another disk on the backup server came from the other one with the same filesystem, duplicate UUID can't be a coincidence, just generate a new one and you're done.

Tried to do that, didnt seem to work. Here's the disk log after the attempt

 

 

[JorgeB]

5 minutes ago, superloopy1 said:

Tried to do that, didnt seem to work. Here's the disk log after the attempt

UD is timing out, try it manually, on the console type:

 

xfs_admin -U generate /dev/sdk1

 

And post any errors.

[superloopy1]

yup ...

 

Log inconsistent or not a log (last==0, first!=1)
empty log check failed
xlog_is_dirty: cannot find log head/tail (xlog_find_tail=22)

ERROR: cannot find log head/tail, run xfs_repair

[JorgeB]

20 minutes ago, superloopy1 said:

run xfs_repair

 

[superloopy1]

Have tried that, in maintenance mode xfs_repair /dev/sdk1 -L (first run said to recreate log) which ended ok.

 

Then back to retry mount, again failed with duplicate uuid, then tried to recreate uuid from 'settings', seemed ok.

 

Back to second attempt to mount and still failing (timing out still) with same error messages as before, duplicate uid, bad superblock et al.

 

Any clever code to try and get around this?

 

Looks like a cmd line to change uuid might be needed?

 

Still sitting, outside of array, wont mount.

Edited February 2, 2020 by superloopy1

[JorgeB]

Run xfs_repair -L manually, then change UUID manually, try try to mount with UD and post error if it still fails.

[superloopy1]

stuck now ....

 

xfs_admin -U  asking for a new uuid?? What do i put here?

[JorgeB]

OK, lets go one step at a time, please post the output of xfs_repair -v /dev/sdX1

[superloopy1]

Just got my head around it now ... found 'uuidgen' was needed. Applied manually and new uuid has now stuck, drive mounted externally ok so hopefully thats the end of it.  Thanks for the help, as you can see i'm no linux guy but am learning, mainly by stuff breaking, every day. I'll make a note of this little episode. Fortunately it was just a copy of my data and not the real thing!! Cheers!!

[JorgeB]

👍