Larger Storage Footprint after Prune and Corruption #1219
Comments
github-actions
bot
added
the
S-triage
This is because of the 2-phase-pruning. Either re-run prune after 24h or (in case you can ensure that no parallel access is made to the repository) use Are you sure you are having corruption? The cache error is strange, but that shouldn't harm anything. |
|
In general, if you have questions about the repository size, |
aawsome
added
C-question
|
Ahh, ok I wasn't aware of the 2-phase pruning, thank you for clarifying. I was worried that there was another issue I wasn't seeing, just didn't want to lose these backups. I should mention that the file size was correct on the original packs there were repaired and I confirmed there was no corruption. Thank you for the quick response. |
I ran prune for the first time in a while after upgrading to v0.8.0 and had some strange behavior start to appear. I received errors from the prune command and then tried to repair index, then repair snapshots, followed by some manually pruning of ungrouped snapshots. I received different errors in most of the commands and my bucket size increased from 4.1GB to 5.2GB after prune, increasing in size after each prune command.
I have a database that is backed up with stdin and it seems that those packs need to have their size corrected when running
repair snapshotbut this is a small database so it shouldn't have a major impact on size. I'm not as worried about this database backup but more so the possibility of corruption of other backups and the unchecked increase in bucket size.Any idea why the bucket would increase in size after running
prune?This is the slimmed down progression of commands and the output.
# NO ERRORS THAT I CAN SEE rustic -P home_bb repair index --log-level debugFull Output
rustic -P home_bb repair snapshots # Many of these errors for different snapshots [INFO] file grafana_db.sql: corrected file size [INFO] saved modified snapshot as d9190e5b. [INFO] file 00023870: corrected file size [INFO] saved modified snapshot as 1313ba9d.Full Output
Strange Error When Running Prune Again
Ran
checkwith no errors, thought everything was ok. Ran interactive snapshots and deleted some old testing snapshots and that were backed up incorrectly, along with the repaired db snapshots because they weren't needed. The snapshots I removed were the largest in the repo 2-3GB each.I ran prune again and it showed an even larger repository size, even though I didn't perform a backup
Also running repair snapshots again led to the same db file backup snapshots being repaired again (different snapshots than the first time but same backup). Any ideas as to why this corruption is happening and why the repo would increase in size after prune?
FYI, the command used to backup the DB is:
The text was updated successfully, but these errors were encountered: