Hi Mailing-Lister's, I am reaching out for assistance regarding a deployment issue I am facing with Ceph on a 4 node RKE2 cluster. We are attempting to deploy Ceph via the rook helm chart, but we are encountering an issue that apparently seems related to a known bug (https://tracker.ceph.com/issues/61597). During the OSD preparation phase, the deployment consistently fails with an IndexError: list index out of range. The logs indicate a problem occurs when configuring new Disks, specifically using /dev/dm-3 as a metadata device. It's important to note that /dev/dm-3 is an LVM on top of an mdadm RAID, which might or might not be contributing to this issue. (I swear, this setup worked already) Here is a snippet of the error from the deployment logs:
2023-11-23 23:11:30.196913 D | exec: IndexError: list index out of range 2023-11-23 23:11:30.236962 C | rookcmd: failed to configure devices: failed to initialize osd: failed ceph-volume report: exit status 1 https://paste.openstack.org/show/bileqRFKbolrBlTqszmC/
We have attempted different configurations, including specifying devices explicitly and using the useAllDevices: true option with a specified metadata device (/dev/dm-3 or the /dev/pv_md0/lv_md0 path). However, the issue persists across multiple configurations. tested configurations are as follows: Explicit device specification: ```yaml nodes: - name: "ceph01.maas" devices: - name: /dev/dm-1 - name: /dev/dm-2 - name: "sdb" config: metadataDevice: "/dev/dm-3" - name: "sdc" config: metadataDevice: "/dev/dm-3" ``` General device specification with metadata device: ```yaml storage: useAllNodes: true useAllDevices: true config: metadataDevice: /dev/dm-3 ``` I would greatly appreciate any insights or recommendations on how to proceed or work around this issue. Is there a halfway decent way to apply the fix or maybe a workaround that we can apply to successfully deploy Ceph in our environment? Kind regards,