Used space gets reset after restart

Alexey · November 23, 2024, 7:53am

Ok. I was able to reproduce in storj-up.

used-space-filewalker doesn't update the usage correctly after restart

opened 07:52AM - 23 Nov 24 UTC

Bug

**Description** After restart of storagenode, the `used-space-filewalker` doesn't update the usage since v1.115.4 See https://forum.storj.io/t/used-space-gets-reset-after-restart/28408?u=alexey **Steps to reproduce the issue:** 1. used `storj-up` environment 2. upload several 128MiB files with TTL +2m (or any) 3. run `docker compose exec storagenode1 storagenode dashboard --address=:30002` to see a dashboard 4. Wait for the TTL collector chore to complete (better several hours) 5. Check the dashboard - it should show some non-zero value (perhaps the TTL collector doesn't update the usage again or not in all databases - we now have an additional database `used_space_per_prefix.db`) 6. upload several 128MiB files with TTL 7. Make sure, that the usage value on the dashboard has changed 8. Restart the node: ``` docker compose exec storagenode1 rm /var/lib/storj/.local/share/storj/storagenode/storage/trash/.trash-uses-day-dirs-indicator && docker compose restart storagenode1 ``` 9. The value would be the same, which was before uploading additional 128MiB files. 10. after the used-space-filewalker has completed the scan, the value will not change. **Describe the results you expected:** The `used-space-filewalker` should calculate the used space correctly **Describe the results you received:** The `used-space-filewalker` was finish in seconds, the used space will be invalid. **Possible Fix** Remove the `used_space_per_prefix.db` database and restart the node, the used space value will be updated correctly. **Your environment** - Operating system and version: any - Additional environment details (Raspberry PI, Docker, VMWare, etc.): any

Just need to keep it running for more than 24h.