Replies: 1 comment 1 reply
-
I am running both RBD and CephFS on Talos via CSI and an external Ceph cluster (Not Rook) Keeping it somewhat brief, here's how I am configured.
Here's an example of creating/mounting a CephFS directory/fs Sorry for the brevity. I really wanted to respond but couldn't spend a ton of time writing up a big beautiful response at the moment. If you'd like, I can meet in discord to help out. |
Beta Was this translation helpful? Give feedback.
1 reply
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
-
Has anyone successfully set this up on Talos? I have an already existing Ceph cluster set up with both CephFS and RBD, but when applying Ceph's suggested helm chart, I get numerous errors. The node plugins seem to be failing due to the immutability of Talos, e.g. logs giving me
"failed to generate container X spec: failed to mkdir "/etc/selinux"
And my PVC is showing "Waiting for a volume go be created either by the external provisioner...".
The only guide to Ceph on the Talos documentation is using Rook, which I cannot do as my cluster is external to k8s.
Relevant links:
Edit: I figured out part of the problem, but not all. My node plugins now seem happy and running when I noticed there was an selinux mount bool in the config and turned that off, but still running into a
failed to provision volume with StorageClass "csi-cephfs-sc": rpc error: code = InvalidArgument desc = volume not found
Edit 2: I gave up on CephFS and tried RBD, which gave me
multi node access modes are only supported on rbd block type volumes
, so I tried setting my PVC toReadWriteOnce
instead ofReadWriteMany
and it seems like it bound?Edit 3: It would be cool if CephFS would work as it means it's easier to mount and access in other spots
Beta Was this translation helpful? Give feedback.
All reactions