]> git.apps.os.sepia.ceph.com Git - ceph-ansible.git/commit
purge: use sysfs to unmap rbd devices
authorGuillaume Abrioux <gabrioux@redhat.com>
Mon, 4 Nov 2019 14:59:39 +0000 (15:59 +0100)
committerDimitri Savineau <savineau.dimitri@gmail.com>
Mon, 13 Jan 2020 19:50:29 +0000 (14:50 -0500)
commit5db0b239f6131a0a40776c2a6c65420db438c057
tree8e97de23737649c9b428b5d075ee21b545377744
parentbcd7fee18dffa52d7752f19ed208e11f92839432
purge: use sysfs to unmap rbd devices

in containerized context, using the binary provided in atomic os won't
work because it's an old version provided by ceph-common based on
10.2.5.
Using a container could be an idea but for large cluster with hundreds
of client nodes, that would require to pull the image of each of them
just to unmap the rbd devices.

Let's use the sysfs method in order to avoid any issue related to ceph
version that is shipped on the host.

Closes: https://bugzilla.redhat.com/show_bug.cgi?id=1766064
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
(cherry picked from commit 3cfcc7a105156dfde65b23e9d8662cd848537094)
infrastructure-playbooks/purge-docker-cluster.yml