Fedora Account System
Red Hat Associate
Red Hat Customer
It was seen in https://bugzilla.redhat.com/show_bug.cgi?id=2276565 (4.15) that client operator upgrade didn't trigger csi-addons upgrade from 4.14 to 4.15, similarly in 4.16 we don't have a clear logical owner who can update csi-addons depending on deployment modes. We want to tie upgrade of csi-addons to the operator that becomes a logical owner depending on deployment mode as follows: 1. only ocs-client-operator deployed: ocs-client-op should upgrade csi-addons 2. odf in non-provider mode: odf-operator should upgrade csi-addons 3. odf in provider mode: ocs-client-op should upgrade csi-addons We need to continue this for atleast upto a couple of releases (~4.18). Testing: 1. Deploy 4.16 in above mentioned cases 2. When the ODF is upgraded to earlier builds of next release (4.17) csi-addons should get upgraded as well. Specifically mentioning earlier builds as by 4.17 FF builds might get changed.
Verified in ODF 4.16.0-125 ODF on provider cluster upgraded from 4.16.0-125 to 4.17.0-19. Ocs-client-operator and odf-csi-addons-operator were installed in the same namespace of odf-operator(this is the default configuration). Ocs-client-operator and odf-csi-addons-operator were upgraded automatically to 4.17.0-19. % oc get csv NAME DISPLAY VERSION REPLACES PHASE mcg-operator.v4.17.0-19.stable NooBaa Operator 4.17.0-19.stable mcg-operator.v4.16.0-125.stable Succeeded metallb-operator.v4.16.0-202405161711 MetalLB Operator 4.16.0-202405161711 metallb-operator.v4.15.0-202405161507 Succeeded ocs-client-operator.v4.17.0-19.stable OpenShift Data Foundation Client 4.17.0-19.stable ocs-client-operator.v4.16.0-125.stable Succeeded ocs-operator.v4.17.0-19.stable OpenShift Container Storage 4.17.0-19.stable ocs-operator.v4.16.0-125.stable Succeeded odf-csi-addons-operator.v4.17.0-19.stable CSI Addons 4.17.0-19.stable odf-csi-addons-operator.v4.16.0-125.stable Succeeded odf-operator.v4.17.0-19.stable OpenShift Data Foundation 4.17.0-19.stable odf-operator.v4.16.0-125.stable Succeeded odf-prometheus-operator.v4.17.0-19.stable Prometheus Operator 4.17.0-19.stable odf-prometheus-operator.v4.16.0-125.stable Succeeded recipe.v4.17.0-19.stable Recipe 4.17.0-19.stable recipe.v4.16.0-125.stable Succeeded rook-ceph-operator.v4.17.0-19.stable Rook-Ceph 4.17.0-19.stable rook-ceph-operator.v4.16.0-125.stable Succeeded OCP version is 4.17.0-0.nightly-2024-06-07-045541 on provider cluster (upgraded from 4.16.0 before the testing) Clients were on hosted clusters using kubevirt and agent(2 each). Total 4 storageclients were connected from 4 hosted clusters. Subscription channel of client-operator in these hosted clusters were changed automatically to stable-4.17. After changing the catalogsource image to 4.17.0-19, ocs-client-operator and odf-csi-addons-operator on client clusters were upgraded automatically to 4.17.0-19. % oc get csv NAME DISPLAY VERSION REPLACES PHASE ocs-client-operator.v4.17.0-19.stable OpenShift Data Foundation Client 4.17.0-19.stable ocs-client-operator.v4.16.0-125.stable Succeeded odf-csi-addons-operator.v4.17.0-19.stable CSI Addons 4.17.0-19.stable odf-csi-addons-operator.v4.16.0-125.stable Succeeded OCP version on kubevirt based hosted clusters is 4.15.13 and agent based hosted clusters is 4.16.0-ec.6.
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 OpenShift Data Foundation 4.16.0 security, enhancement & bug fix update), 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-2024:4591