The --force flag would allow the command to delete its data directly, it wouldn’t move it to the trash.
The trash is collected by a Garbage Collector, if the data was deleted by customers, and it would be deleted automatically after 7 days in the trash.
I read that somewhere in the docs, yet my trash level is at 50-55% and has been like that for months. I don’t think that is normal, is it?
Can I remove it directly from the file system maybe?
Yes, you can remove from the trash subfolder date-named subfolders older than 7 days in each satellites subfolder (see Satellite info (Address, ID, Blobs folder, Hex) for the info). Please note, you shouldn’t delete from the blobs subfolders of the trusted satellites, otherwise node will be disqualified.
However then you would need to restart the node with the enabled scan on startup and wait until used-space-filewalker will finish calculations without errors and updated the databases, this will correct the showed usage on the pie chart of the dashboard.
1 Like
Looks like Storj node is thinking there is trash, but actually there isn’t (according to file system utilization). Restarted with --storage2.piece-scan-on-startup, hopefully it fixes itself now.
1 Like
Well, it has been running with --storage2.piece-scan-on-startup for ~36 hours. Trash is still 10.89TB/20TB with zero indications of making any progress anywhere. I also have 0B of free space and 0.89TB of Overused even though I have had 20TB allocated for months and there was no overuse just 36 hours ago.
Something is seriously messed up in Storj node implementation of space accounting, but I don’t know what or now to debug it.
if you use lazy filewalker, what is default setup, then it will take very long time.
2 Likes
You can check timestamp of older folders to check if files inside them are getting deleted.
1 Like
I only see folders from last 7 days, which is in-line with expectations.
Restarted with both --storage2.piece-scan-on-startup and --pieces.enable-lazy-filewalker=false, now it seems to be doing something, we’ll see.
In that case I would also suggest to enable a badger cache as well. The first run would be slow as usual, but any next one will be faster.
Okay, after almost 2 days of running, it finally finished.
221GB of trash, 0B overused, finally looking sane.
Thanks everyone!
2 Likes
Good job my friend 