i’ve noticed lots of files in my repositories index and data subdirectories that are not covered by the every-file-is-named-by-its-content-hash-policy.
the naming scheme is the following:
one example would be:
(not matching SHA256 of content)
the repository was created about 2 years ago (v1).
those files appear since about 1 year, according to their timestamps.
the repository is periodically pruned, but those files stay.
what are they about and in case they are really unneeded (“temp”), why doesn’t a prune get rid o fthem?
to partly answer my question:
they are part of the SFTP backend upload procedure.
since their hashes don’t match and they were not removed, i guess the connection dropped during upload.
the question why they did not get pruned, still remains.
Removing backup snapshots — restic 0.15.2 documentation.
Simple, there’s no cleanup implemented for these files so far. Feel free to open an issue . How many files are we talking about?
there were about 50-100 of those files.
i’ve opened an issue, as i expected a prune to remove all superfluous files: prune does not remove all superfluous files · Issue #4372 · restic/restic · GitHub