Multi-node with MAAS, machine is missing disks
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
OpenStack Snap |
Triaged
|
Undecided
|
Unassigned |
Bug Description
Running the openstack 2024.1/edge snap for a multi-node with MAAS deployment results in:
Error: Machine 'microceph/0' is missing disks: {'/dev/
Currently the manifest is deploying microceph reef/edge, rev 32.
If I manually list the disks, the microceph snap responds appropriately:
$ juju ssh -m openstack-machines microceph/0 "sudo microceph disk list --json"
{"ConfiguredDis
Connection to 192.168.100.140 closed.
If I run the microceph charm list-disks action, it returns empty:
$ juju run -m openstack-machines microceph/0 list-disks
Running operation 13 with 1 task
- task 14 on unit-microceph-0
Waiting for task 14...
osds: '[]'
unpartitioned-
{'osds': [], 'unpartitioned-
If I change the manifest to use the microceph latest/edge (r47) charm things seem to work as expected:
$ juju run -m openstack-machines microceph/0 list-disks
Running operation 21 with 1 task
- task 22 on unit-microceph-0
Waiting for task 22...
osds: '[{''osd'': 1, ''path'': ''/dev/
''location'': ''lab02''}, {''osd'': 2, ''path'': ''/dev/
''location'': ''lab03''}, {''osd'': 3, ''path'': ''/dev/
''location'': ''lab04''}]'
unpartitioned-
Seems to be similar symtpoms as https:/ /bugs.launchpad .net/snap- openstack/ +bug/2065470.
Need to wait for next reef/candidate release.