Description of problem: When an OSD service spec contains container_args, the OSDs deploy as normal but then they will redeploy signalling that a container_arg change has been seen - when it hasn't. This causes new OSDs to redeploy and restart and depending on the number of OSDs in the service spec, this could affect 00's of OSDs and cluster performance. Version-Release number of selected component (if applicable): 19.2.0-55 How reproducible: All the time :) Steps to Reproduce: 1. Deploy an osd service spec with container arguments 2. Watch the cephadm log for the deployment, followed by a redeploy 3. Actual results: OSDs get restarted unecessarily - in the lab environment this added 1 hour to the cluster expansion process Expected results: When there isn't a container args change, the OSDs should not be redeployed/restarted Additional info: This has been discussed on Slack with Adam, who has confirmed the behavior locally.
Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory (Important: Red Hat Ceph Storage 8.1 security, bug fix, and enhancement updates), and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. https://access.redhat.com/errata/RHSA-2025:9775