August 6, 20232 yr Oh boy - I am freaking out. Since yesterday my Unraid shows the following: Alle array Disks (xfs) are unmountable: unsupported partition layout. The single drive Cachepool (btrfs) is unmountable: unsupported or no fs This looks really bad - I am close to a heartattack, pls help me. hsunraid-diagnostics-20230806-1141.zip
August 6, 20232 yr Community Expert The devices have no partitions, so something or someone deleted them, you can try recreating them as they were before, they need to have the exact same settings, and the data should come back.
August 6, 20232 yr Author How do recreate them ? I have no Raid or s.th. like it for the nvme 😞 How can that happen Edited August 6, 20232 yr by Schulmeister
August 6, 20232 yr Community Expert For the array, and assuming parity is valid, easiest way is to rebuild one disk at a time, since Unraid will recreate the correct partition, to test, stop the array, unassign one of the data disks, start the array, if the emulated disk mounts and contents look correct re-assign the disk to rebuild on top, then repeat for the other disks (one by one). For cache it will need to be manually re-created, but since it's 4TB it will be GPT so cannot test at the moment as I'm about to leave for the day, try the array first and then post back.
August 6, 20232 yr Author It might work with the array drives - takes a day per disk ! Is there any way to speed this up ?
August 6, 20232 yr Community Expert 21 minutes ago, Schulmeister said: It might work with the array drives Do you mean you get a mountable emulated disk when you do this? 28 minutes ago, JorgeB said: stop the array, unassign one of the data disks, start the array, if the emulated disk mounts and contents look correct
August 6, 20232 yr Author I unassigned Disk1 and got to see the files and folder again. Then I reassigned the Disk again and it showed a warning saying emulated content The Data rebuild takes roughly a day per disk if I do it like this - with seven Disks that means I have a week bevore me.
August 6, 20232 yr Author Regarding the single cachepool NVME, I have read a lot on btrfs and gdisk I startet it with: gdisk /dev/nvme1n1 with this result: - what should I do ??
August 6, 20232 yr Author I am copying an image right now (takes some 4 hrs). Is gdisk the way to go ? It seems the Backup GPT and Part-Table are OK
August 6, 20232 yr Author I have rebuild the disk using gdisk. My heart literaly stoped when I pressed w (rite) - well I had a GPT Backup and a Full Disk to Image running before. Now it shows this: My Data-Rebuild on Disk1 is still running, so I am not brave enough to reboot just now. It will take another 10-12 hrs für the first disk - after that I try rebooting the server. Wish me luck
August 7, 20232 yr Author Good morning Jorge, after a reboot the nvme still shows "unsupported or no filesystem" Any ideas ? I am so cunfused right now, I don't know if this was a btrfs or a zfs Disk ? Any way to find out for sure ? Edited August 7, 20232 yr by Schulmeister
August 7, 20232 yr Community Expert The partition is OK, but it's not the only problem, parts of the filesystem were deleted/damaged: Aug 7 12:12:34 HSUnraid kernel: BTRFS error (device nvme1n1p1): bad tree block start, mirror 1 want 23625728 have 0 Aug 7 12:12:34 HSUnraid kernel: BTRFS error (device nvme1n1p1): bad tree block start, mirror 2 want 23625728 have 0 You can try the recovery options here, especially btrfs restore, but this error can be fatal.
August 7, 20232 yr Community Expert Forgot to mention, for the other array disks, since they are mostly 16TB you should be able to copy the current partition from disk1 and restore to the others (except the 18TB disk7), if you want try one first: With array stopped: sfdisk -d /dev/sdX > part_table replace X with disk1 sfdisk /dev/sdY < part_table replace Y with other disk. Then reboot and see if it's back, if yes, do it for the others, except disk7, if not do the rebuild thing.
August 7, 20232 yr Author I tried: root@HSUnraid:/mnt/disk1# btrfs restore -vi /dev/nvme1n1 /mnt/disk1/restore No valid Btrfs found on /dev/nvme1n1 Could not open root, trying backup super No valid Btrfs found on /dev/nvme1n1 Could not open root, trying backup super No valid Btrfs found on /dev/nvme1n1 Could not open root, trying backup super As I mentioned before - I am not really sure if the nvme was btrfs, is there a way to check it - to be certain
August 7, 20232 yr Author 24 minutes ago, JorgeB said: Forgot to mention, for the other array disks, since they are mostly 16TB you should be able to copy the current partition from disk1 and restore to the others (except the 18TB disk7), if you want try one first: With array stopped: sfdisk -d /dev/sdX > part_table replace X with disk1 sfdisk /dev/sdY < part_table replace Y with other disk. Then reboot and see if it's back, if yes, do it for the others, except disk7, if not do the rebuild thing. That did not work 😞
August 7, 20232 yr Community Expert 34 minutes ago, Schulmeister said: As I mentioned before - I am not really sure if the nvme was btrfs It was, but like mentioned that error can be fatal. 23 minutes ago, Schulmeister said: That did not work 😞 The partition is back, but there's fs corruption: Aug 7 14:04:46 HSUnraid kernel: XFS (md2p1): Metadata CRC error detected at xfs_agi_read_verify+0x85/0xfa [xfs], xfs_agi block 0x2 Aug 7 14:04:46 HSUnraid kernel: XFS (md2p1): Unmount and run xfs_repair To play it safer unassign disk2 and see if UD plugin can repair the filesystem, that way you still have the rebuild option to try if it can't.
August 7, 20232 yr Community Expert Just click the bolt check mark type icon after unassigning the device form the array.
August 7, 20232 yr Author I did unassign it but there is no bolt-type Icon - neither near Disk2 or UD section ? I have a mount Button at the UD section - do you mean the gears ? Edited August 7, 20232 yr by Schulmeister
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.