Page MenuHomeDevCentral

Claim space on Stormshear datastore
Closed, ResolvedPublic

Description

Storage on Stormshear

When we created extra partition for ''arcology'' ZFS pool on Ysul and ''wharf'' LVM group on Dwellers, we have underestimated the overhead of the VMs.

ServerTaskDescriptionSize
Ysul-Main hard disk400 G
YsulT150ZFS arcology100 G
Dwellers-Main hard disk100 G
DwellersT166Wharf for Docker200 G
Total..800 G

But we have some overhead. For example, it seems Ysul has snapshots enabled (I think this were a safety measure to handle T293), as we have this:
-rw------- 1 root root 94.3G Aug 2 13:07 Ysul-000001-delta.vmdk
-rw------- 1 root root 3.5G Apr 8 19:08 Ysul-Snapshot1.vmsn
-rw------- 1 root root 7.8G Aug 2 13:03 Ysul-zfspool-000001-delta.vmdk

Ysul arcology

We don't need so much space, as we currently use it for jails and poudriere.

$ zpool list
NAME       SIZE  ALLOC   FREE   FRAG  EXPANDSZ    CAP  DEDUP  HEALTH  ALTROOT
arcology  99.5G  17.0G  82.5G      -         -    17%  1.00x  ONLINE  -

Plans

  • Destroy the 100G arcology ZFS pool and recreate it on a 25G storage
  • Disable snapshot on Ysul - I totally know it's a good idea as it's also a dev machine, and so not totally instrumentable, but we don't have the hard disk to do that. A proper backup strategy and a Tahoe grid is needed.

Event Timeline

dereckson claimed this task.
dereckson raised the priority of this task from to High.
dereckson updated the task description. (Show Details)
dereckson added a project: Servers.
dereckson added a subscriber: dereckson.

I'm deleting Ysul snapshot and consolidating it into the machine.

Operation is at 19%.

More than 100 GB have been recovered.

$ ssh root@stormshear.nasqueron.org df -h
Filesystem   Size   Used Available Use% Mounted on
VMFS-5     924.0G 808.0G    116.0G  87% /vmfs/volumes/datastore1
[...]

Situation is stable since August, so we can consider space has been successfully reclaimed.