Hey, I'm looking to figure out why on the Main tab in Unraid, appdata (unassigned disk) is reporting as 240GB used (1GB more than this morning), when it was about 50GB smaller yesterday. The first thing that comes to mind is that I needed to re-create my docker.img yesterday, but I'm not seeing the old one hiding anywhere.

 

I've attached three images - what Unraid is reporting the drive usage as, and what Krusader and qdirstat are reporting the drive usage as.

 

 

 

 

Edited February 6, 20242 yr by ctrlaltd1337

The Unraid dashboard reports based on "df -H". It's better for monitoring if you're going to run out of space; as such, the more useful column there is the "Free" column. I don't use Krusader, but I bet it's using du to calculate disk usage. They use different techniques for reporting: df reports what the file system says it has; du loops over the directories and adds up the file sizes. Because of this, df is fast, du is slow (you might be able to notice this when you use Krusader to give you disk utilization).

 

As for why they're different: du doesn't run on the whole file system; if the process running du doesn't have sufficient permissions, it can't see everything (since it's manually looping over the directories); if the files are deleted, but still open by another process (which happens a lot with log files), du can't see them but df "can" (in the sense that it's still reported by the file system), etc.

 

Keep in mind we're also looking at GB vs GiB here.

 

Oversimplified tl;dr: df (and the Unraid stats page) are good for telling you if you're going to run out of room (free disk); du is better for telling you what is using that storage.

2 hours ago, ctrlaltd1337 said:

Hey, I'm looking to figure out why on the Main tab in Unraid, appdata (unassigned disk) is reporting as 240GB used (1GB more than this morning), when it was about 50GB smaller yesterday. The first thing that comes to mind is that I needed to re-create my docker.img yesterday, but I'm not seeing the old one hiding anywhere.

 

Sorry, I realized I missed your main question. 

 

Have you run a trim recently? Try that first 'fstrim -va'.

 

It's also possible (for the reasons in my previous post) that you have deleted files being held open by a process. These files are seen by df but not by du. There are a couple of commands you can try here to figure out what:

 

'lsof +aL1 /unassigned/ud-docker'

'lsof -n | grep -i deleted'

 

Of course, you can also just try restarting Unraid if you haven't already.

The first command didn't show any results, and when I restarted my server, the unassigned disk didn't auto-mount. And when I mounted it, my Docker image failed to start again ("Docker Service failed to start." when trying to open Docker tab). 😭 But the used space has gone down to 188GB used out of 250GB again.

 

I'm starting to wonder if it's my SSD, but it didn't throw any errors and the SMART looks to be fine?

 

 

 

So Docker is down now? That's a different issue. Turn off Docker under Settings (and apply). Reboot Unraid. Make sure unassigned disk has mounted. Re-enable Docker.

Yeah, following those steps worked to re-instate my Docker image, and without issue.

 

The size on the disk continues to fluctuate, currently at 209GB. I'll see if the varying in size continues and I'll change my appdata to a proper cache pool and mirror it soon as well. Thanks for talking through this with me, I'll have to see if it happens again!