-
Bug
-
Resolution: Unresolved
-
Critical
-
None
-
None
-
Quality / Stability / Reliability
-
3
-
False
-
-
False
-
ToDo
-
-
-
Very Likely
-
0
-
None
-
Unset
-
Unknown
-
None
Upstream Issue: https://github.com/vmware-tanzu/velero/issues/9294
What steps did you take and what happened:
With Velero 1.17, I tried to use fs-backup on OpenShift. I get a permission error because the fs-backup pod is not running in a Privileged security context, which is required for fs-backup to work on OpenShift. In 1.16, this worked because we have the option to run the node agent pods in privileged mode.
What did you expect to happen:
The backup to work
The following information will help us better understand what's going on:
I'll be submitting a PR shortly. We need a new --privileged-fs-backup install and node-agent server option for users using k8s environments that require this. We need a new param rather than simply propagating the existing node-agent privileged flag, since the conditions where this is needed for node agent (block mode on all clusters) are different from where it's needed for fs-backup (for certain cluster types if using fs-backup). It's possible to need privileged node agent but not privileged fs-backup, and vice versa.
Environment:
- Velero version: 1.17
- Kubernetes platform: OpenShift
- Issue affects filesystem backup functionality requiring privileged pod security context
Additional Notes:
This issue affects OpenShift clusters specifically where filesystem backup requires privileged security context to access mount paths. The functionality worked in Velero 1.16 but is broken in 1.17.