mgr/cephadm: enable use_repo_digest by default for pacific

[sebastian-philipp]

Let's keep the cluster consistent by default.

Signed-off-by: Sebastian Wagner sebastian.wagner@suse.com

Checklist

• References tracker ticket
• Updates documentation if necessary
• Includes tests for new functionality or reproducer for bug

---

Show available Jenkins commands

• jenkins retest this please
• jenkins test classic perf
• jenkins test crimson perf
• jenkins test signed
• jenkins test make check
• jenkins test make check arm64
• jenkins test submodules
• jenkins test dashboard
• jenkins test api
• jenkins test docs
• jenkins render docs
• jenkins test ceph-volume all
• jenkins test ceph-volume tox

[sebastian-philipp]

pytest failed:

>                   assert image == 'to_image'
E                   AssertionError: assert 'to_image@repo_digest' == 'to_image'
E                     - to_image
E                     + to_image@repo_digest

cephadm/tests/test_upgrade.py:89: AssertionError

[liewegas]

Hmm, I wonder if instead of making this a global option like this, it would make more sense to do something like

ceph orch upgrade start --image ceph/daemon-base:master --pin (or --pin-to-digest)

to pin it to a specific digest/version, instead of an ambiguous/floating tag.

[sebastian-philipp]

Hmm, I wonder if instead of making this a global option like this, it would make more sense to do something like

ceph orch upgrade start --image ceph/daemon-base:master --pin (or --pin-to-digest)

to pin it to a specific digest/version, instead of an ambiguous/floating tag.

doesn't work.
Imagine:

1. specify ceph/daemon-base:master as the base image which is at this point equal to 16.2.1
2. cephadm deploys mon.a and mon.b
3. someone pushes 16.2.2 to ceph/daemon-base:master
4. Now future daemon deployments might deploy 16.2.2 without the user knowing about that.

we have to use the digest consistently for all future deployments, independent of an upgrade or not.

[liewegas]

Imagine:

1. specify ceph/daemon-base:master as the base image which is at this point equal to 16.2.1
2. cephadm deploys mon.a and mon.b
3. someone pushes 16.2.2 to ceph/daemon-base:master
4. Now future daemon deployments might deploy 16.2.2 without the user knowing about that.

we have to use the digest consistently for all future deployments, independent of an upgrade or not.

When doing an upgrade to ceph/daemon-base:master, with --pin we'd translate that to ceph/daemon-base@something in container_image, so new daemons would have the same version. We could do something similar during bootstrap for the base case install.

[github-actions]

This pull request can no longer be automatically merged: a rebase is needed and changes have to be manually resolved

[github-actions]

This pull request can no longer be automatically merged: a rebase is needed and changes have to be manually resolved

[sebastian-philipp]

• https://pulpito.ceph.com/swagner-2021-02-10_11:41:39-rados:cephadm-wip-swagner-testing-2021-02-09-1126-distro-basic-smithi/
• https://pulpito.ceph.com/swagner-2021-02-11_10:31:33-rados:cephadm-wip-swagner-testing-2021-02-09-1126-distro-basic-smithi/

Failures:

• https://tracker.ceph.com/issues/49259
• e2e test

[sebastian-philipp]

https://tracker.ceph.com/issues/50502