Hi, I am running 26.0.0-BETA.1 and I have two ppols: apps and datapool. Datapool is spinning rust and has severl TB still available. The pool apps is running on 2 mirrored SSD (440 GiB) and it shows 370 GiB used, 70 GiB free. The only visible dataset under apps are my configs for the installed apps at below 3 GiB. There are no snapshots activated and there are no shown when clicking in GUI for snapshots. I see the disc usage creeping up and have not idea where the 370 GiB are used for and if I can reduce the usage and get rid of the warning…
Any pointers for me?
root@truenas[/mnt/apps]# du -h --max-depth=1
3.0G ./configs
32K ./ix-applications
I can probably try to get rid of the ./ix-applications, but 32K is not the issue here
No, it isn’t. The issue is almost certainly the ix-apps dataset, which exists on that pool, but it’s mounted in a different location (specifically, /mnt/.ix-apps), so du won’t see it. But if you check the output of zfs list, it’ll show up, and I’d bet that’s where the 370 GB are.
ok, zfs list did the trick and you were right. some config folders are quite big in ix-apps (jellyfin, nextcloud) and it looks like I have to upgrade. sucks (while looking at the current cost level of ssd )
You could try to move that data to the spinners; depending on what the data is you may not lose much performance. But are you sure it’s just config? I store my media on spinners, and the Jellyfin metadata/config on the SSD pool; the latter runs to about 30 GB, and I have a fairly large library. I can’t imagine hitting hundreds of GB in just metadata/config/database files–though if the Nextcloud data is also there, that would explain it pretty easily.
)