Help requested to recover corrupt repository

Support
1

Hi everyone! I wonder if you can help me out.
My drive crashed; and I have troubles with both backups: 1) clonezilla/partclone cloned image of btrfs partition failed to restore (corrupt after restore) so I’m down to: 2) kopia backups not accessible (stored on NAS NFS).

The repository NFS is mounted with options: rw,noatime,vers=4.1,hard,proto=tcp)
On kopia repository connect filesystem --path /share/Kopia-Beast --progress --log-level debug :

Creating cache directory /share/Data-Archive/BeastKopia/cache' with max size 5242880000
(Finished initial cache scan...)
failed to open repository: unable to create shared content manager : error loading indexes: unable to load pack indexes despite 10 retries

14:11:33.363 total active indexes 973, deletion watermark 2025-09-14 23:39:09 +0700 +07
14:11:33.366 Downloading 973 new index blobs...
14:11:33.366 encountered NOT_FOUND when loading, sleeping 6.4s before retrying #7
14:11:39.767 adding incremental state for epochs 0..2 on top of []
14:11:39.767 Complete Index Set for [0..2]: [xn0_005fc01fd36426053746b5a2d4b1910b-s6a61385b60709484139-c1 xn0_0102ce46bd64f7db50d3382e7d773ce2-s37698ce84cebb263139-c1 xn0_0153ae7ac44139df6d28257649e4f245-s01738ec957133606139-c1 xn0_016b527d6ae3cf6ae2b666b59bc21779-sba2756c7f797199b138-c1 xn0_01cf473cc5139906a01624d404a872ce-s24681b26c74ec88c139-c1 xn0_01dc36069b71c8569d980bd3dd58f31c-sa6f6092da41569f2139-c1 
[...]

Are there advanced/hidden maintenance commands in Kopia to rebuild indexes from the files alone?
Version 0.22.0 (same error with 0.21 and 0.20)

I’ve posted the same message to Slack community; I can remove it from either one as needed.

2

I hope you manage to salvage some of your data.

But only true solution to such problems is to follow 3-2-1 backup rule:) Things happen. Disks die, cloud providers can go offline and repositories can get corrupted beyond repair. Long term it is always not if but when.

3

Hi kapitein thank you for responding!

Yes I had a btrfs corrupt clonezilla clone from a few weeks back but I managed to recover it in the end with dmde.

I might still have lost some DB data across services so in still looking for a way to restore my kopie backups. Do you know if its possible to create a new repo, rsync the broken repo files, and then reindex?

4

Sorry no. I do consider any such attempts too risky for my data so never bothered to investigate. If my repo gets corrupted (super rare thing) I delete it and start fresh. Then copy historical data from another repo.

Or for S3 cloud with object lock configured I simply roll back to some older repo version (not corrupted). But this is theory as it never happened - yet.

5

Maybe have a look at this thread, where someone resolved a similar issue:

https://kopia.discourse.group/t/kopia-repository-corrupted-itself-unable-to-add-to-xxx-index-blob-cache-context-canceled/1766