Hello,

Due do bad sata cables and an unclean shutdown, I had 3 disk marked "unmountable".
I did repair the file system on each three, it was talking about bag log but they never got mountable. After zero log the disks :
- Disk 1 has all its data in lost&found, in random named folder
- Disk 2 has just a few files, nothing alarming here
- Disk 5 has all its data in lost&found, in random named folder

Disk 1 & 2 support my Nextcloud share, sporting 4 TB of data.
Disk 5 & 6 are versioned backups of the nextcloud share, done with Rsync.

I just rebuilt disk 5 on a brand new disk but I have the older one (although giving a lot of read errors).

In the Lost & Found folders, files seems okay with proper names, but put randomly.
 

Any ideas ?

 

Edited April 17, 20251 yr by xoC
Sent too soon.

I am afraid that sorting out a lost+found is basically a manual job, and often it is not worth the effort involved.

 

If you do want to try and sort it then the Linux 'file' command can be used to at lest give you the content type of files so you know what application can be used to examine its contents.

If I put back the old disk5, unraid will force a rebuild, but it concerns only data, not file structure right ?
So maybe I can get something working by reinstalling the old disk5, getting back the old file structure, and I can recover from one of my versioned backups ?

Edit : I should mention that I repaired the file system of disk 5 after swapping disks and data rebuild on the new disk.
So the old one is probably unmountable, but I don't think it was "zero log-ed".

Edited April 17, 20251 yr by xoC

All the files I'm concerned about are the ones stored in the Nextcloud share.
I've stopped Nextcloud for the time being, but does Nextcloud rely on the file system of the array, or it relies on its own database for that and I could recover from Nextcloud ?

Edited April 17, 20251 yr by xoC

With further analysis :
Disk 1 has 4 folders in lost & found :
- 3 of them contain a few files
- the last one seem to contain all the data of the Nextcloud share, properly ordered. And the size / number of file seems coherent.


Could it be possible that the file system error concerned just the root folders of the disk, and everything else was preserved ?
If I move that folder from lost&found to the Nextcloud share, should it work ?

Edit : is it safe to say that the three other folders are very old lost & founds since their modified date is 2021 ? In green the folder with the nextcloud share.

Sorry for the gazillion questions, I'm slightly panicking.

Edited April 17, 20251 yr by xoC

So after checking in detail every disk.
1) My data from the nextcloud share is properly ordered at once, just in a Lost&Found Folder.
2) My backup data on disk 5 is a total mess, split in thousand+ folders. This is backup data from my nextcloud share, so if I can get back my nextcloud share as it was, I can rebuild new backup from it instead of trying to order all backup data on disk 5 (which are split between disk 5 & 6, only data on 5 is a mess).

2) is not a problem if I can recover from 1).

now big question I didn't find the answer looking on the net :
How can I move the data from Disk 2 to Disk 2 (renaming the nextcloud share folder and moving it in the root folder), and be sure that all timestamps will be preserved ?
It seems mv can't guarantee that.
cp can but I don't have enough space on the disk to do a copy preserving timestamp then deleting original.

Any ideas ?
 

3 hours ago, xoC said:

If I put back the old disk5, unraid will force a rebuild, but it concerns only data, not file structure right ?
So maybe I can get something working by reinstalling the old disk5, getting back the old file structure, and I can recover from one of my versioned backups ?

Don't really understand what you mean by "only data, not file structure".

 

In any case, the only way to put original disk5 back in the array with its contents intact is New Config and rebuild parity.

 

You could try to mount original disk5 as an Unassigned Device and see if you can make any sense of its contents.

44 minutes ago, xoC said:

How can I move the data from Disk 2 to Disk 2 (renaming the nextcloud share folder and moving it in the root folder), and be sure that all timestamps will be preserved ?

Entries in lost+found is where the repair process could not find the directory entry for a file/folder it found.

 

You have not indicated the current path on disk2 and the path you want to end up with, but you should be able to do something like

mv /mnt/disk2/lost+found/foldername /mnt/disk2/NextCloud

adjusting path names as appropriate and assuming that NextCloud is the name of the share you want to end up with.

 

If you do that then the timestamp of the /mnt/disk2/NextCloud folder will probably not be maintained, but I think all the containing files/folders will be.  Just for interest why are you that worried about the timestamps anyway - surely getting the files back in the correct location is the key requirement and the timestamps is just a nicety?

I tried the mv command on one of the old folder in lost+found and... it doesn't change the timestamps (at least the one of the last modification, which i'm interested in) !  

Last modified timestamps are important for me, as my nextcloud share host my professional data (I'm a sound engineer) and my software checks that timestamp for every subproject that is rendered and imported in a main project. If the date doesn't match, when opening an old project (for referencing), it will force a re-rendering of everything. And since software/plugins florish and die fast, that would break a lot of projects.

Thanks for your help !