Actions
Bug #63998
openContainers have too old lsblk. Partitions aren't reported correctly
Status:
New
Priority:
Normal
Assignee:
-
Category:
container tools
Target version:
-
% Done:
0%
Source:
Tags:
Backport:
Regression:
No
Severity:
3 - minor
Reviewed:
Affected Versions:
ceph-qa-suite:
Pull request ID:
Crash signature (v1):
Crash signature (v2):
Description
If OSD nodes have logical drives and I want Ceph to use the partition of the drives it won't work. It's because lsblk that comes with the container doesn't report the partitions as partitions and therefore ceph-volume ignores them. Could the lsblk be upgraded to a version that reports the partition correctly? Also, if there is a workaround I would be very intereted in learning it.
Here is an example:
# grep PRETTY /etc/os-release PRETTY_NAME="Red Hat Enterprise Linux 9.3 (Plow)" # cephadm version cephadm version 18.2.1 (7fe91d5d5842e04be3b4f514d6dd990c54b29c76) reef (stable) # cephadm --image d2cdd87030d1 inspect-image { "ceph_version": "ceph version 18.2.1 (7fe91d5d5842e04be3b4f514d6dd990c54b29c76) reef (stable)", "image_id": "d2cdd87030d19393788f48f06e1a453e100160debff5a32c6b537a2b2e450e8e", "repo_digests": [ "quay.io/ceph/ceph@sha256:5e15da7c3173070a55d3cafdf76075fbd4e49f348f1efec53235290f85659c6e" ] } # lsblk | grep md └─md0 9:0 0 1.5T 0 raid1 ├─md0p1 259:4 0 1T 0 part └─md0p2 259:5 0 466.3G 0 part └─md0 9:0 0 1.5T 0 raid1 ├─md0p1 259:4 0 1T 0 part └─md0p2 259:5 0 466.3G 0 part # cephadm shell Using ceph image with id 'd2cdd87030d1' and tag '<none>' created on 2023-12-15 19:31:03 +0000 UTC quay.io/ceph/ceph@sha256:5e15da7c3173070a55d3cafdf76075fbd4e49f348f1efec53235290f85659c6e # lsblk -V lsblk from util-linux 2.32.1 # lsblk | grep md `-md0 9:0 0 1.5T 0 raid1 |-md0p1 259:4 0 1T 0 md `-md0p2 259:5 0 466.3G 0 md `-md0 9:0 0 1.5T 0 raid1 |-md0p1 259:4 0 1T 0 md `-md0p2 259:5 0 466.3G 0 md
Updated by Matti Saarinen 4 months ago
One essential information I forgot. The host has lsblk 2.37.4.
Actions