Jottacloud - an update

So a while ago (four months) I migrated from Hetzner to Jotta - it’s been smooth sailing until last month, when I got a bunch of errors thrown at me.

Initial checking led me to believe something had gone haywire on the Jotta side - some files could no longer be downloaded, nor through the web ui or rclone. Not a development I’m super thrilled about… it did give me some opportunity to see how well restic can deal with this.

I spun up a VM on Hetzner, copied over my configuration, and started running restic check --read-data. That came back with errors.

In summary:

restic repair index to repair the index
restic repair packs <packId> to repair whichever packs it could repair
restic repair snapshots --forget to clean up the snapshots

Using the troubleshooting info from the official documentation and this post on the restic forums for some pointers on how to deal with broken trees, everything is back to working order.

I did modify my backup script a little:
It was running a check of 1.5% of the data each day which I’ve modified now check 1/<amount of days in the month>’s worth of data, using the day of the month as the section it needs to verify.

You can easily calculate this using the following shell command:

SLICE="$(date +%d)/$(date -d "$(date +%Y-%m-01) + 1 month - 1 day" +%d)"

I also added a bandwidth limit to rclone for uploads using --bwlimit 8M:off. This has vastly reduced the amount of retries I get from rclone.