openstack overcloud ceph deploy fails

Bug #1986903 reported by Cristian Le
6
This bug affects 1 person
Affects Status Importance Assigned to Milestone
tripleo
New
Undecided
Unassigned

Bug Description

First issue:
- `Error: container-init binary not found on the host: stat /usr/libexec/podman/catatonit: no such file or directory`
- Installing `catatonit` seems to resolve this issue

Second issue:
- `error={"changed": true, "cmd": ["podman", "run", "--rm", "--net=host", "--ipc=host", "--volume", "/etc/ceph:/etc/ceph:z", "--volume", "/home/ceph-admin/assimilate_ceph.conf:/home/assimilate_ceph.conf:z", "--volume", "/home/ceph-admin/specs/ceph_spec.yaml:/home/ceph_spec.yaml:z", "--entrypoint", "ceph", "quay.rdoproject.org/tripleomastercentos9/daemon:current-ceph", "--fsid", "c417e52c-3561-589e-a46d-71c6278929f1", "-c", "/etc/ceph/ceph.conf", "-k", "/etc/ceph/ceph.client.admin.keyring", "orch", "apply", "--in-file", "/home/ceph_spec.yaml"], "delta": "0:00:00.647363", "end": "2022-08-18 00:39:01.006893", "msg": "non-zero return code", "rc": 22, "start": "2022-08-18 00:39:00.359530", "stderr": "Error EINVAL: Failed to validate OSD spec \"default_drive_group\": `placement` required", "stderr_lines": ["Error EINVAL: Failed to validate OSD spec \"default_drive_group\": `placement` required"], "stdout": "", "stdout_lines": []}`
- ~~This one I don't know how to debug~~. The latter is because there was no ceph storage node defined
- There are no checks here to confirm that ceph storage and ceph controllers are present in `baremetal_deployed.yaml`

Third issue:
- There is no test if there are 3 ceph storage nodes configured. Without it, the ceph deployment gets stuck at `Wait for expected number of osds to be running`

Fourth issue:
- The playbook does not check if a default route gateway is properly defined
- I have found that it is necessary to have this setup

Final issue:
- There is a ceph version mismatch between the image built `cephadm` and containers, i.e. pacific vs quincy
- Manually installing `centos-ceph-release-quincy` and `dnf upgrade` resolves this issue

Cristian Le (lecris)
description: updated
Cristian Le (lecris)
description: updated
Cristian Le (lecris)
description: updated
To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.