I think I have a trash accounting problem

[Alexey]

@jammerdan

And now the trash deleter cannot finish, because it will not delete this pieces and errors the folder is not empty.

Do you have an evidence, that it didn’t continue to delete other folders?
I also think that corrupted piece may appear in the trash folder itself, not only in a blobs folder.

@jammerdan

Well the question is, will it resume where it was as we have the save state resume feature or will it break on error not save and start from beginning. This is a huge difference.

It should. But probably it has a bug:

Used space gets reset after restart

Hello, Is it normal that the used space drops after a reboot? my node is a couple of days old, and i see in the dashboard 108gigs used space First time i restarted the vm at around 40gigs, used dropped to ±7gigs Now i restarted again, used dropped from 108gig to 8.2gig trash and overused are both 0 but if i do df -h i see that there is 167gig in use there are so many logentries but i dont really see errors at first node is running in a dedicated storj proxmox vm and with no additional dr…

@jammerdan

So there are no free space updating that process?

It should update the usage when it’s finished the batch deletion and flushed to the database after 1 hour by default. However, if there was an interruption in the middle of the process, I think the progress could be lost and you would have a discrepancy there, which should be fixed by the used-space-filewalker enabled on startup. For your system I would recommend to never disable the scan on startup, because seems your setup is unstable.