Best Practice - Backup - Check - Prune

Here is my schedule. All tasks are once per day

Backup at 12:00
4 Jobs not run in parallel
Check at 1:00 (This takes 5.5 hours until the GCD fix is incorporated in a release)
Prune at 6:00

Does this make sense?

Why not put all steps into the same schedule?

It depends on how often your data changes and prune schedule that may be driven by your storage provider cost structure. Generally I’d think backup once a day is too seldom while check and prune once a day is way too often.

As an example, on my desktop i personally run backup every 20 min, check every month and prune every week.

For my B2 backups I would love to reduce the frequency of my check schedule, but I’m using Duplicacy Web-UI for handling my cloud backups and so far it seems that if a storage has not had a recent (or recent-enough) check completed then the dashboard view becomes completely useless. Is that some known issue? Is there a workaround that I’m not familiar with? Right now the only way I can get a working Web-UI dashboard is to have a 1:1 ratio between backups and checks.

Can you elaborate on that? Do plots and stats show wrong information? Don’t show any information? Can you post a screenshot?

Here is a screenshot of my fairly recent web-ui instance (running on a Synology NAS, backing up to another Synology NAS – but the backend should not matter) with check twice a week and hourly backup.

Each graph updates at its own cadence and reflects current information. Do you see something else with back blaze?

Yes, if I stop running nightly check jobs, then both graphs just stop updating. They will have gaps in their data until I run a check job, then immediately after check succeeds the graphs will “magically” fill in with the missing data. I’ll update my check schedule to only run weekly and then in a few days after I have a bad-looking dashboard graph I’ll post a screenshot.

I have noticed something similar to @roland.cassini 's issue. I am also on B2.


I have check set to run manually only. I did a check on May 31, and just did one today, June 29. In between those dates Storage Size is reported as zero; a correct graph should show a slowly rising storage size from about 80 GB to about 90 GB over this time period. In my case, the Storage Size graph did not fill in with correct data after I did a check today; just the endpoints, when I did the two checks, are correct.

And in my case, the New Revisions graph is plausibly correct.