Description of problem: I tried to upgrade openshift-kube-descheduler from 4.7.2 to latest but i fail to upgrade and upon further investigation it looks like descheduler operator skiprange needs to be updated. current skipRange is '>=4.3.0-0 < 4.7.0-0 where as it should be >=4.3.0 < 4.7.0 For more info, we could refer to logging one: https://github.com/openshift/cluster-logging-operator/blob/release-4.5/manifests/4.5/cluster-logging.v4.5.0.clusterserviceversion.yaml Also discussion related to this: https://coreos.slack.com/archives/CH76YSYSC/p1616568447464200 Version-Release number of selected component (if applicable): [knarra@knarra ~]$ oc get csv -n openshift-kube-descheduler-operator NAME DISPLAY VERSION REPLACES PHASE clusterkubedescheduleroperator.4.7.0-202101300133.p0 Kube Descheduler Operator 4.7.0-202101300133.p0 Succeeded How reproducible: Always Steps to Reproduce: 1. Install 4.7.2 cluster 2. Deploy descheduler operator 3. Edit the catalogSource index image and set that to latest Actual results: Could see that operator does not get upgraded to latest version of 4.7.z Expected results: Operator should get updated to latest version of 4.7.z Additional info:
I opened a PR against master at https://github.com/openshift/cluster-kube-descheduler-operator/pull/177. Let me know if this needs to be backported to the 4.7 branch I'm also lowering the severity, because from the linked slack discussion it sounds like this only applies to development upgrades, and the ART tooling automatically replaces this pattern for releases. Correct me if I misunderstood that
Hi Mike, Moving the bug back to assigned state as all other operators have the skiprange as below. So i think descheduler as well should be inline with them. https://github.com/kube-reporting/metering-operator/blob/master/manifests/deploy/o[…]m/bundle/4.8/meteringoperator.v4.8.0.clusterserviceversion.yaml https://github.com/openshift/cluster-nfd-operator/blob/master/manifests/olm-catalog/4.8/nfd.v4.8.0.clusterserviceversion.yaml#L40 https://github.com/openshift/ocs-operator/blob/master/deploy/bundle/manifests/ocs-operator.clusterserviceversion.yaml#L994 https://github.com/openshift/compliance-operator/blob/master/deploy/olm-catalog/co[…]erator/manifests/compliance-operator.clusterserviceversion.yaml If you look at the above operators yaml i see the skipRange to be olm.skipRange: ">=4.5.0 <4.8.0" and not ">4.5.0-0 <4.8.0" Could you please help take a look again ? Thanks !!
@Rama, that is the change that was merged in https://github.com/openshift/cluster-kube-descheduler-operator/pull/177/files, our skipRange is now defined as ">=4.3.0-0 < 4.7.0" which matches those. Please double check, as that should be merged into the 4.7 and 4.8 descheduler operator manifests now.
@mike should not it be ">=4.3.0 < 4.7.0" ? When i looked at other operator manifests that is how they are. Could you please help clarify if this should be fine ?
Ah, I see what you mean (I missed the hyphen in the lower version number). Opened https://github.com/openshift/cluster-kube-descheduler-operator/pull/184 to fix that too
Verified the bug with the build below and i see that i could successfully upgrade from 4.7 to latest 4.8 , the skipversion looks to be correct as per the PR. [knarra@knarra ~]$ oc get csv -n openshift-kube-descheduler-operator NAME DISPLAY VERSION REPLACES PHASE clusterkubedescheduleroperator.4.8.0-202104231341.p0 Kube Descheduler Operator 4.8.0-202104231341.p0 clusterkubedescheduleroperator.4.7.0-202104090228.p0 Succeeded [knarra@knarra ~]$ oc get csv clusterkubedescheduleroperator.4.8.0-202104231341.p0 -n openshift-kube-descheduler-operator -o yaml| grep skip olm.skipRange: '>=4.6.0 < 4.8.0' f:olm.skipRange: {} will clone a bug for 4.7.z
I could successfully upgrade from 4.7 to 4.8 but the skipRange still does not correspond to art.yaml's search pattern, so moving it back to assigned.
Verified with the build below and i could successfully upgrade from 4.7 to 4.8 and now i see that the skipRange correspond to art.yaml's search pattern. [knarra@knarra new_way]$ oc get csv NAME DISPLAY VERSION REPLACES PHASE clusterkubedescheduleroperator.4.8.0-202104292205.p0 Kube Descheduler Operator 4.8.0-202104292205.p0 clusterkubedescheduleroperator.4.7.0-202102110027.p0 Succeeded [knarra@knarra new_way]$ oc get clusterversion NAME VERSION AVAILABLE PROGRESSING SINCE STATUS version 4.8.0-0.nightly-2021-04-30-201824 True False 89m Cluster version is 4.8.0-0.nightly-2021-04-30-201824 [knarra@knarra new_way]$ oc get csv -o json | grep skip "olm.skipRange": "\u003e=4.6.0 \u003c4.8.0-202104292205.p0", "f:olm.skipRange": {}, Based on the above moving bug to verified state.
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 (Moderate: OpenShift Container Platform 4.8.2 bug fix and security 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-2021:2438