Opened 3 months ago

Closed 3 months ago

Last modified 3 months ago

#3291 closed defect (fixed)

osgeo3 gallery VM frozen

Reported by: darkblueb Owned by: sac-tickets@…
Priority: normal Milestone: Sysadmin Contract 2024-I
Component: SysAdmin Keywords: osgeo3, gallery
Cc:

Description

accidental disk full from Image Data (tree study on HuggingFace)

I was logged in and deleted some large files when I noticed the condition, but the VM is not responding at this time.

Maybe a simple reboot can do it? or restore to the last snapshot is OK

(first trouble in years, apologies for the accident) -Brian

Change History (5)

comment:1 by robe, 3 months ago

I rebooted but space did not clear up.

I deleted all the snapshots and created a new snapshot. That only cleared 26G of space. See if you can clear more space. I don't want to increase the size cause it's already 1.4T and will be hard to shrink if I increase it.

comment:2 by darkblueb, 3 months ago

.. maybe something is odd with this VM.. the $df -h tool reports that the VM disk space is osgeo3_lxd/containers/gallery 869G .. not at all 1.4T

.. deleting a 2GB file did not change the free space reported by $df -h .. ? not sure how that is possible

comment:3 by robe, 3 months ago

Milestone: UnplannedSysadmin Contract 2024-I

comment:4 by robe, 3 months ago

Resolution: fixed
Status: newclosed

It's a zfs partition so the snapshots get overlaid as diffs. So when you clear disk space disk space is not actually cleared until snapshots are deleted. I just deleted all the daily snapshot created today and created a new snapshot (so it's pretty much a 0 byte snapshot).

You should now see 1.4T again with 557G available.

comment:5 by darkblueb, 3 months ago

confirm osgeo3_lxd/containers/gallery 1.4T 841G 556G 61% /

Note: See TracTickets for help on using tickets.