[Bug]: Truncated "filecache" table, now data is wrongly linked

[huntervcx]

[Bug]: Truncated "filecache" table, now data is wrongly linked

[nickvergessen]

Tried to truncate oc_filecache multiple times

Congratulations, you just broke your server. You should not purge that table. The fileids are reused in other places, e.g. in shares and that is mostlikely what is causing your issue

[huntervcx]

Issue was before the truncate on a new server initialized with the same smb shares; and same issue on other server on v29.0.0.19 (with auto truncate enabled, admit for this one)
Tried the truncate as it was a same behavior as #45238 (showing wrong folder listing, truncate was the only way to refresh)

Problem remains the same : NC doesn't seems to load the files from backend share but only from his filecache table. And doesn't compare file_hash, as it opens files from other shares, that the user never had access to

[solracsf]

Why in hell did you think it was a good idea to truncate an essential table for Nextcloud? This could NEVER be a solution for anything 😭

Only a backup restore can save you, otherwise you're good to restart your server from scratch.

Sorry to say, but this is not a bug, and no-one can help you here. 🥴

[huntervcx]

Issue was before the truncate on a new server initialized with the same smb shares; and same issue on other server on v29.0.0.19 (with auto truncate enabled, admit for this one) Tried the truncate as it was a same behavior as #45238 (showing wrong folder listing, truncate was the only way to refresh)

Problem remains the same : NC doesn't seems to load the files from backend share but only from his filecache table. And doesn't compare file_hash, as it opens files from other shares, that the user never had access to

As said just before, on a new server with same setup, same behavior after a while (>20days) and random folders/files, and pretty big oc_filecache table with fileid corresponding on other tables.
truncate was a test on a secondary server, with same config, that temporarly fixed other issues of listing folders on external storage

the issue is here without any modification to the db. That's why I reported it initally, as it's random and sometimes revert to the normal files after a lot of time (>6hours while being under massive load from users)
Issue also here with or without redis for cache and filelock, with and without preview enabled, some stock apps enabled and disabled

[solracsf]

In that case continue your debug at #45238 because here there is nothing we can do about it.

AGAIN, oc_filecache is not a "cache" in any means (even if the name of the table is not perfect, it should be more like oc_filetable).