You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
This is a bit of a broad ask but can we have the ODF must-gather custom resources restructured to be more inline with how OCP must-gathers structure this data. For example, OCP must-gathers have the following structure for CRs:
<mgroot>/<scope>/<CRgroup>/<CRKind>/<CR.yaml>
If we take the VolumeSnapshotContent resources as an example. In an OCP must-gather they're located here:
But in an ODF must-gather, it seems we have CRs in arbitrary locations. Just to name a couple, our storagecluster and storagesystem CRs are in the "oc_output' directory:
$ ls namespaces/openshift-storage/oc_output/
all_-o_wide csv events_get hpa pods rolebinding storagecluster storageconsumer.yaml
configmap events_desc events_yaml installplan role secrets storagecluster.yaml storagesystem.yaml
And the cephcluster CRs are in the ceph/namespaces/openshift-storage/ceph.rook.io/ceph directory (I concede this is fairly similar to where an OCP must-gather would put this, but the root directory is "ceph" and not the root of the ODF must-gather:
$ ls ceph/namespaces/openshift-storage/ceph.rook.io/ceph
cephblockpools/ cephclusters/ cephfilesystems/ cephobjectstores/ cephobjectstoreusers/
The reason for this ask (besides obvious clarity and coherency) is that Red Hat support has just gained awareness of [1] which extends the OMC tool to include CRDs. We're unable to take full advantage of this extension because many of the ODF CRs are not in the location that the omc tool expects them to be in.
I am going to keep this open. I love the idea where once can use a cli to inspect the collected must-gather. It would benefit both engineering and support.
It would require significant refactoring but would be worth the effort.
This is a bit of a broad ask but can we have the ODF must-gather custom resources restructured to be more inline with how OCP must-gathers structure this data. For example, OCP must-gathers have the following structure for CRs:
<mgroot>/<scope>/<CRgroup>/<CRKind>/<CR.yaml>
If we take the VolumeSnapshotContent resources as an example. In an OCP must-gather they're located here:
<mgroot>/cluster-scoped-resources/snapshot.storage.k8s.io/volumesnapshotcontents/
But in an ODF must-gather, it seems we have CRs in arbitrary locations. Just to name a couple, our storagecluster and storagesystem CRs are in the "oc_output' directory:
And the cephcluster CRs are in the
ceph/namespaces/openshift-storage/ceph.rook.io/ceph
directory (I concede this is fairly similar to where an OCP must-gather would put this, but the root directory is "ceph" and not the root of the ODF must-gather:The reason for this ask (besides obvious clarity and coherency) is that Red Hat support has just gained awareness of [1] which extends the OMC tool to include CRDs. We're unable to take full advantage of this extension because many of the ODF CRs are not in the location that the omc tool expects them to be in.
[1] gmeghnag/omc#56
The text was updated successfully, but these errors were encountered: