restic: Restic can fail with `Fatal: failed to refresh lock in time` on slow connections

Output of restic version

0.15.1

Expected behavior

Backups should also be possible over a slow network connection.

Actual behavior

When uploading over a slow network connection, it is possible that a backup or similar operation fails to refresh lock files within 22.5 minutes and fails with Fatal: failed to refresh lock in time. This has been reported on IRC and at least at https://github.com/restic/restic/issues/4157#issuecomment-1408669114 .

Steps to reproduce the behavior

Use a very slow network, then wait for half an hour…

Do you have any idea what may have caused this?

restic only allows a limited number of concurrent backend connections. Once all are used for uploading pack files, then a lock file refresh is queued behind them. For a slow connection, it is quite possible that a 16MB pack file (with 5 parallel connections) takes longer than half an hour to upload.

Further fallout from #3569.

Do you have an idea how to solve the issue?

Lock file refreshes should be able to bypass the connections limit.

About this issue

  • Original URL
  • State: closed
  • Created a year ago
  • Reactions: 3
  • Comments: 25 (7 by maintainers)

Most upvoted comments

We are also observing this same error. Restic v 0.15.1 fails on a slow connection with a rather small backup of about 600MB after about 22 minutes… I would also request you to treat this as a bug please.

Unfortunately I have the same issue, I can’t backup on some pc because the connection is slow. The only workaround is to use 0.14.0 version.

IMHO this has to be marked as a bug: if a backup program can’t backup data in some conditions this is bug, it is its main purpose.