deploy storage with constraints on lxd error

[Erik Lönroth]

When I deploy a charm on my lxd requesting a root disk constraint, it fails with not much helpful messages.

Here is my charm code: https://github.com/erik78se/ops-scp/blob/master/scp/README.md

Juju version: 2.9.18-ubuntu-amd64 (snap)
Controller version: 2.9.16 and 2.9.18
lxd cloud: 4.0.8 21835 4.0/stable (snap)

===========
juju deploy ./scp_ubuntu-20.04-amd64.charm --constraints "cores=2 mem=1G root-disk=5G"
Located local charm "scp", revision 25
Deploying "scp" from local charm "scp", revision 25
========

Model Controller Cloud/Region Version SLA Timestamp
rafaelbackpssh iceberg-default iceberg/default 2.9.18 unsupported 00:15:37+01:00

App Version Status Scale Charm Store Channel Rev OS Message
scp waiting 0/1 scp local 25 ubuntu waiting for machine

Unit Workload Agent Machine Public address Ports Message
scp/13 waiting allocating 18 waiting for machine

Machine State DNS Inst id Series AZ Message
18 down pending focal Failed creating instance record: Failed initialising instance: Failed loading storage pool: No such object

If I deploy without the "root-disk" constraint - it works just fine. Its only when I use the "root-disk" constraint. Do I have something mis-configured in my storage-pools?

$ juju list-storage-pools
Name Provider Attributes
loop loop
lxd lxd
lxd-btrfs lxd driver=btrfs lxd-pool=juju-btrfs
lxd-zfs lxd driver=zfs lxd-pool=juju-zfs zfs.pool_name=juju-lxd
rootfs rootfs
tmpfs tmpfs

==== On the lxd-remote =======

$ lxc storage list iceberg:
+------------+--------+-----------------------------------------------+-------------+---------+
| NAME | DRIVER | SOURCE | DESCRIPTION | USED BY |
+------------+--------+-----------------------------------------------+-------------+---------+
| juju | dir | /var/snap/lxd/common/lxd/storage-pools/juju | | 0 |
+------------+--------+-----------------------------------------------+-------------+---------+
| juju-btrfs | btrfs | /var/snap/lxd/common/lxd/disks/juju-btrfs.img | | 0 |
+------------+--------+-----------------------------------------------+-------------+---------+
| juju-zfs | zfs | /var/snap/lxd/common/lxd/disks/juju-zfs.img | | 0 |
+------------+--------+-----------------------------------------------+-------------+---------+
| lxdhosts | zfs | lxdhosts | | 18 |
+------------+--------+-----------------------------------------------+-------------+---------+

==== On the lxd-remote =======
$ zpool list
NAME SIZE ALLOC FREE CKPOINT EXPANDSZ FRAG CAP DEDUP HEALTH ALTROOT
juju-zfs 80M 136K 79.9M - - 5% 0% 1.00x ONLINE -
lxdhosts 556G 229G 327G - - 38% 41% 1.00x ONLINE -

[Erik Lönroth (erik-lonroth)]

Here is a juju status

[Harry Pidcock (hpidcock)]

Hi Erik,

It looks like because you don't have a storage pool named "default".

Try:
juju deploy ./scp_ubuntu-20.04-amd64.charm --constraints "cores=2 mem=1G root-disk=5G root-disk-source=lxdhosts"

I think this is still a valid error since juju should have made an attempt to either use the juju-zfs/juju-btrfs pool or correctly query lxd for the default pool set in the profile.

[Erik Lönroth (erik-lonroth)]

I've tried to deploy as:

juju deploy ./scp_ubuntu-20.04-amd64.charm --storage data=lxd,2G,1 --constraints "cores=2 mem=1G root-disk=5G root-disk-source=lxdhosts"

This is how it looks inside the container after deploy. It looks like I can't get a 2G disk from this either.

Filesystem Size Used Avail Use% Mounted on
lxdhosts/containers/juju-a9629a-23 5.9G 1.3G 4.6G 22% /
none 492K 4.0K 488K 1% /dev
udev 189G 0 189G 0% /dev/tty
tmpfs 100K 0 100K 0% /dev/lxd
tmpfs 100K 0 100K 0% /dev/.lxd-mounts
tmpfs 189G 0 189G 0% /dev/shm
tmpfs 38G 212K 38G 1% /run
tmpfs 5.0M 0 5.0M 0% /run/lock
tmpfs 189G 0 189G 0% /sys/fs/cgroup
/dev/mapper/ubuntu--vg-ubuntu--lv 134G 18G 110G 14% /srv/scponly/data
tmpfs 1.0M 0 1.0M 0% /var/snap/lxd/common/ns
tmpfs 103M 0 103M 0% /run/user/1000

Is it impossible to limit disk quota with lxd/zfs/juju?

[Harry Pidcock (hpidcock)]

No it should be applying the limit in the backend.

Try this on the host:
$ sudo zfs list lxdhosts/containers/juju-a9629a-23

Otherwise maybe there is a minimum of 5.9gb, which seems strange.

[Harry Pidcock (hpidcock)]

When I tried with 2G root disk I get

$ sudo zfs list -o name,quota juju-zfs/containers/juju-b49df8-3
NAME QUOTA
juju-zfs/containers/juju-b49df8-3 2G