Hide Forgot
Description of problem: After creating a subscription, the PHASE of the csv is empty. [jzhang@dhcp-140-18 installer]$ oc get csv -n jian NAME DISPLAY VERSION REPLACES PHASE etcdoperator.v0.9.2 etcd 0.9.2 etcdoperator.v0.9.0 Check the olm-operator logs, got below errors: E1109 06:02:06.241510 1 reflector.go:205] github.com/operator-framework/operator-lifecycle-manager/pkg/lib/queueinformer/queueinformer_operator.go:108: Failed to list *v1alpha2.OperatorGroup: the server could not find the requested resource (get operatorgroups.operators.coreos.com) Check the catalog-operator logs, got below: time="2018-11-09T05:58:51Z" level=info msg="jian/install-etcdoperator.v0.9.2-99hnn updated" time="2018-11-09T05:58:51Z" level=info msg="error fetching CSV etcdoperator.v0.9.2 via k8s api: clusterserviceversions.operators.coreos.com \"etcdoperator.v0.9.2\" not found" I confirm that the "etcdoperator.v0.9.2" csv exists in the ConfigMap. packages: " \ \n- #! package-manifest: ./deploy/chart/catalog_resources/rh-operators/etcdoperator.v0.9.2.clusterserviceversion.yaml\n \ packageName: etcd\n channels:\n - name: alpha\n currentCSV: etcdoperator.v0.9.2\n Version-Release number of selected component (if applicable): image: registry.svc.ci.openshift.org/openshift/origin-v4.0-20181109054332@sha256:bd42b78b320f01a3c5b9181d5581b77a9103c4bcf1e9c4d7bbfe1c1b098b6ba8 [jzhang@dhcp-140-18 installer]$ oc exec olm-operator-67f6b68849-tglt5 -- olm -version OLM version: 0.7.4 git commit: c82dca5 How reproducible: always Steps to Reproduce: 1. Install the OCP 4.0 via openshift-installer by following the doc: https://github.com/openshift/installer/blob/master/docs/dev/libvirt-howto.md 2. Create a subscription, like below: [jzhang@dhcp-140-18 installer]$ cat subscription.yaml apiVersion: operators.coreos.com/v1alpha1 kind: Subscription metadata: generateName: etcd- namespace: jian spec: source: rh-operators name: etcd startingCSV: etcdoperator.v0.9.2 channel: alpha [jzhang@dhcp-140-18 installer]$ oc create -f subscription.yaml 3. Check the Etcd pod in project "jian". [jzhang@dhcp-140-18 installer]$ oc get subscription -n jian NAME PACKAGE SOURCE CHANNEL etcd-7r662 etcd rh-operators alpha [jzhang@dhcp-140-18 installer]$ oc get installplan -n jian NAME CSV SOURCE APPROVAL APPROVED install-etcdoperator.v0.9.2-99hnn etcdoperator.v0.9.2 rh-operators Automatic false [jzhang@dhcp-140-18 installer]$ oc get csv -n jian NAME DISPLAY VERSION REPLACES PHASE etcdoperator.v0.9.2 etcd 0.9.2 etcdoperator.v0.9.0 [jzhang@dhcp-140-18 installer]$ oc get pods -n jian No resources found. Actual results: The status of the csv was incorrect, and no pod generated. [jzhang@dhcp-140-18 installer]$ oc get csv -n jian NAME DISPLAY VERSION REPLACES PHASE etcdoperator.v0.9.2 etcd 0.9.2 etcdoperator.v0.9.0 [jzhang@dhcp-140-18 installer]$ oc get pods -n jian No resources found. Expected results: Subscription successfully. No errors and the Etcd pod can be generated. Additional info:
When I test TSB and ASB operator, I also hit this issue, subscription can not create successfully. olm-operator log: E1119 08:08:26.899521 1 reflector.go:205] github.com/operator-framework/operator-lifecycle-manager/pkg/lib/queueinformer/queueinformer_operator.go:108: Failed to list *v1alpha2.OperatorGroup: the server could not find the requested resource (get operatorgroups.operators.coreos.com) OLM version: 0.8.0 git commit: ff2325e
I used the latest version for having a try. It works well. So, remove the TestBlocker label. [jzhang@dhcp-140-18 installer]$ oc exec olm-operator-585c885c55-2zcr9 -- olm -version OLM version: 0.8.0 git commit: 08ea39b Actually, the version is "git commit: 08ea39b", I highly suggest that we fix bug 1626434 first so that the users can know the version clearly. @Evan What do you think? For now, you can move this bug to "ON_QA" status so that I can verify it. It's better if you can post the fixed PR in here. [jzhang@dhcp-140-18 installer]$ cat subscription.yaml apiVersion: operators.coreos.com/v1alpha1 kind: Subscription metadata: generateName: etcd- spec: source: rh-operators name: etcd startingCSV: etcdoperator.v0.9.2 channel: alpha [jzhang@dhcp-140-18 installer]$ oc get pods NAME READY STATUS RESTARTS AGE catalog-operator-75c8fcc485-md5xn 1/1 Running 0 6h etcd-operator-7b49974f5b-gdd5m 3/3 Running 0 8m olm-operator-585c885c55-2zcr9 1/1 Running 0 6h package-server-7d685ccff7-6sb9j 1/1 Running 0 6h [jzhang@dhcp-140-18 installer]$ oc get csv NAME DISPLAY VERSION REPLACES PHASE etcdoperator.v0.9.2 etcd 0.9.2 etcdoperator.v0.9.0 Succeeded [jzhang@dhcp-140-18 installer]$ oc get installplan NAME CSV SOURCE APPROVAL APPROVED install-etcdoperator.v0.9.2-sx4kp etcdoperator.v0.9.2 rh-operators Automatic false [jzhang@dhcp-140-18 installer]$ oc get subscription NAME PACKAGE SOURCE CHANNEL etcd-kxcgh etcd rh-operators alpha
Bug 1626434, comment 20 outlines a suitable way for QE to keep track of what upstream code was built. I think this bug can be moved to ON_QA.
This is fixed in the latest version.
Verify it per comment 2.
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, 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/RHBA-2019:0758