]> git.apps.os.sepia.ceph.com Git - ceph.git/commit
cephadm: Support Docker Live Restore 59933/head
authorMichal Nasiadka <mnasiadka@gmail.com>
Wed, 11 Sep 2024 12:26:37 +0000 (14:26 +0200)
committerAdam King <adking@redhat.com>
Mon, 23 Sep 2024 15:45:27 +0000 (11:45 -0400)
commitf882fa82fa1bac9d1c8743e2b0d2f654cf9afbb7
tree3a0fa724876c011cbb4416a133b668038be383db
parent1108cc993b0ae058a58aa267d18f607a05fe0e31
cephadm: Support Docker Live Restore

Currently with Docker Live Restore [1] enabled and while restarting
Docker Engine - all Ceph container images will get restarted,
while the feature allows restarting docker.service without
containers downtime.

This is due to Requires=docker.service in systemd units templates,
which mandates that on docker.service restart - the ceph container
systemd units will be restarted as well.

Reworking Requires= to Wants= that is a weaker version of the former,
see [2].

Leaving After= entries, because they should allow systemd to correctly
order the startup (first docker, then ceph containers).

[1]: https://docs.docker.com/engine/daemon/live-restore/
[2]: https://www.freedesktop.org/software/systemd/man/latest/systemd.unit.html#Wants=

Fixes: https://tracker.ceph.com/issues/68028
Signed-off-by: Michal Nasiadka <mnasiadka@gmail.com>
(cherry picked from commit e0f77686523337f37e3ddbbe40eaa92c68947195)
doc/cephadm/install.rst
src/cephadm/cephadmlib/templates/ceph.service.j2
src/cephadm/cephadmlib/templates/init_ctr.service.j2
src/cephadm/cephadmlib/templates/sidecar.service.j2
src/cephadm/tests/test_unit_file.py