Hello Alexey,
Don’t you know how much time I need to scan my mismatched info database with real state? I mean 3.7 TB + 3.7 TB + 8.1 TB allocated space for Storj. I have 6 TB occupied space by Trash for maybe a month. Will turning off lazy filewalker with turning on piece scan on startup will fix my problem? Or i need only to make turn on piece scan on startup and leave lazy filewalker turned on? I would be thankful for your reply.
Best regards.
First of all you need to search for errors related to a databases and/or filewalker in your logs.
You need to search for error and database, error and filewalker.
If you have database related errors like malformed, not a database, then you need to check them for errors and fix them:
If you have database related errors like database is locked, then you need to optimize your disk subsystem or move the databases to the other less loaded disk/SSD.
if you have a filewalker related errors, then you need to optimize your disk subsystem or disable a lazy mode, enable the scan on startup if you disabled it (it’s enabled by default), save the config and restart the node.
With the disabled lazy mode you cannot track the progress via logs, but you may use a debug port (Guide to debug my storage node, uplink, s3 gateway, satellite) with endpoint /mon/ps or check what’s folder is currently processed by the OS tools.
I cannot say how much time it would take on your system to scan your disk(s), only you can tell us. For example, on my biggest node (7TB) (i7 4770 32GB RAM) it was started on 2024-06-28T15:00:11Z and finished for all satellites only on 2024-07-09T07:37:05Z in a lazy mode.
I also had a 7TB lazy one take around 8 or 9 days to finish. (it made my reported space much closer to my OS reported space. Still maybe a couple hundred gigs off).
Another user with closer to 12TB reported 10 days.
It’s very slow.
A “non lazy” (false) filewalker may be a bit faster because it runs with a higher I/O priority. But still slow.
But first check Alexey’s advice and the logs and make sure the filewalker isn’t erroring out.
A post was merged into an existing topic: Fatal Error on my Node
