Hi,
I have the following problem (version doesn’t matter, tested from >=0.17.0). When using multiple clients (Windows, Mac) via KopiaUI and the server directly (e.g., home dockerdata, etc.), it unfortunately often happens that Kopia “blocks” (no progress on either the client or the server). On the server, I create a separate directory for each snapshot. I was able to alleviate it somewhat by running the snapshot creation process on the server at different times. I start Kopia via Cronicle.
The calls look something like this:
docker exec -e KOPIA_CACHE_DIRECTORY=“/app/cache/client/${NAME}” -e KOPIA_LOG_DIR=“/app/logs/client/${NAME}” kopia kopia snapshot create “/Source/${SNAPSHOT}”
.
As a developer, I would suspect a deadlock. Only restarting the server and restarting the snapshot will fix it.
Does anyone have a similar problem?
What exactly do you mean by “no progress”? KopiaUI unresponsive? Does the Tasks tab in KopiaUI list any active tasks? If yes, can you click on it and see details? Do you see activity on the tasks detail page? Do you see any CPU, disk or network activity? Have you tried running Kopia without Cronicle? Kopia should handle scheduling just fine.
All clients are responding. I use KopiaUI on Windows/MacOS via Policy (every 1 hour), and on the SteamDeck, I start it manually. At first, I let Kopia create the backups itself (via Policy), but my backups kept increasing, so I started looking for a graphical “cron” because I wanted to have complex execution times. Now I’ve timed everything so that the jobs don’t overlap, and it’s been working for about a week. But that’s not a solution for me when the backup times increase from almost 2 minutes to almost 45 minutes without any significant changes to the data (etc., opt… <1 minute). Even the Docker data (1.6 TB is normally backed up in about 3 minutes).
If you need some information, logs etc. I can show for them
Increasing backup times sounds like backups are started while another backup is still running. Watch the number of processess while the backup times are increasing. Ideally you would disable Cronicle to check if your issue still appears without it.
Kopia has multiple options to set execution times including cron expressions, which allows multiple entries. This should cover most requirements.
if the problem happens again (I count the number of kopia processs), but now I changed the Backup-strategie, so that overlapping only that the client/server have a lot of data to backup.
Maybe check the “backend” of your Kopia server…? I do have three servers, which are snapshotting up to 450 shares every night and I am running up to 15 jobs simultaneously. Sure, the load on the repo server increased drastically, but I never experienced any lock-ups unless there was an issue with the repo itself. The jobs usually fnish in about an hour or so, unless there is some real data to transfer.