Hide Forgot
Description of problem: On OCP 4.6 cluster, there are 2 duplicated running pods for certified-operators/community-operators/redhat-marketplace/redhat-operators in the project openshift-marketplace. [root@preserve-olm-env daily-test]# oc get pod -n openshift-marketplace NAME READY STATUS RESTARTS AGE certified-operators-mwwcb 1/1 Running 0 12m certified-operators-wgj9f 1/1 Running 0 67m community-operators-hmt56 1/1 Running 0 27m community-operators-m6jr4 1/1 Running 0 12m marketplace-operator-554c465f85-94zrj 1/1 Running 0 72m qe-app-registry-5q6zm 1/1 Running 0 74m redhat-marketplace-bjxlx 1/1 Running 0 12m redhat-marketplace-k2n7k 1/1 Running 0 74m redhat-operators-45ztg 1/1 Running 0 12m redhat-operators-jzqms 1/1 Running 0 74m Version-Release number of selected component (if applicable): [root@preserve-olm-env daily-test]# oc get clusterversion NAME VERSION AVAILABLE PROGRESSING SINCE STATUS version 4.6.0-0.nightly-2020-08-05-082458 True False 83m Cluster version is 4.6.0-0.nightly-2020-08-05-082458 [root@preserve-olm-env daily-test]# oc exec catalog-operator-5d748db9c6-5pt47 -n openshift-operator-lifecycle-manager -- olm --version OLM version: 0.16.0 git commit: 163608d60f37cc3496736bfc4ec72ca01dc7083a How reproducible: Always Steps to Reproduce: 1. Install OCP 4.6 cluster. 2. Check the pods in the project openshift-marketplace. [root@preserve-olm-env daily-test]# oc get pod -n openshift-marketplace NAME READY STATUS RESTARTS AGE certified-operators-mwwcb 1/1 Running 0 12m certified-operators-wgj9f 1/1 Running 0 67m community-operators-hmt56 1/1 Running 0 27m community-operators-m6jr4 1/1 Running 0 12m marketplace-operator-554c465f85-94zrj 1/1 Running 0 72m qe-app-registry-5q6zm 1/1 Running 0 74m redhat-marketplace-bjxlx 1/1 Running 0 12m redhat-marketplace-k2n7k 1/1 Running 0 74m redhat-operators-45ztg 1/1 Running 0 12m redhat-operators-jzqms 1/1 Running 0 74m Actual results: There are 2 duplicated running pods for certified-operators/community-operators/redhat-marketplace/redhat-operators in the project openshift-marketplace. Expected results: There should be 1 running pods for certified-operators/community-operators/redhat-marketplace/redhat-operators in the project openshift-marketplace. Additional info: To debug, the cluster access info is https://mastern-jenkins-csb-openshift-qe.cloud.paas.psi.redhat.com/job/Launch%20Environment%20Flexy/105277/artifact/workdir/install-dir/auth/kubeconfig/*view*/.
Then create subscription in the project test-operators. No InstallPlan, csv or pods created. [root@preserve-olm-env daily-test]# oc create ns test-operators namespace/test-operators created [root@preserve-olm-env daily-test]# oc create -f - <<EOF > apiVersion: operators.coreos.com/v1 > kind: OperatorGroup > metadata: > name: test-operators-og > namespace: test-operators > spec: > targetNamespaces: > - test-operators > EOF operatorgroup.operators.coreos.com/test-operators-og created [root@preserve-olm-env daily-test]# oc create -f - <<EOF > apiVersion: operators.coreos.com/v1alpha1 > kind: Subscription > metadata: > generation: 1 > name: amq-streams > namespace: test-operators > spec: > channel: stable > installPlanApproval: Automatic > name: amq-streams > source: redhat-operators > sourceNamespace: openshift-marketplace > startingCSV: amqstreams.v1.2.0 > EOF subscription.operators.coreos.com/amq-streams created [root@preserve-olm-env daily-test]# oc get sub -n test-operators NAME PACKAGE SOURCE CHANNEL amq-streams amq-streams redhat-operators stable [root@preserve-olm-env daily-test]# oc get ip -n test-operators No resources found in test-operators namespace. [root@preserve-olm-env daily-test]# oc get csv -n test-operators No resources found in test-operators namespace.
this is expected behavior as part of the polling feature of these catalogsources. The old pods should be cleaned up during the next sync cycle. That there is no installPlan is a separate issue that has to do with the use of a stale registry image
OK. How long is the sync interval? Can it be set in the configuration file or by other method? The old pods had been running for more than 1 hours, not be cleaned up. This behavior is strange.
*** Bug 1870367 has been marked as a duplicate of this bug. ***
Version: [hui@localhost verification-tests]$ oc get clusterversion NAME VERSION AVAILABLE PROGRESSING SINCE STATUS version 4.6.0-0.ci-2020-09-12-172918 True False 36h Cluster version is 4.6.0-0.ci-2020-09-12-172918 [hui@localhost verification-tests]$ oc exec olm-operator-67d84fff6c-ndgnc -n openshift-operator-lifecycle-manager -- olm --version OLM version: 0.16.1 git commit: 6d26c16166b232561132985e1132fce4b4d36532 Steps to test: 1, Install OCP 4.6 cluster. 2, Check the catalogsource pods in the project openshift-marketplace. [hui@localhost verification-tests]$ oc get pod -n openshift-marketplace NAME READY STATUS RESTARTS AGE certified-operators-jz4pt 1/1 Running 0 7m53s certified-operators-r7g7n 1/1 Running 0 16h community-operators-krrb6 1/1 Running 0 7m53s community-operators-tgsc7 1/1 Running 0 116m marketplace-operator-54789d9cf4-8cd5h 1/1 Running 0 36h qe-app-registry-gvcj8 1/1 Running 0 36h qe-app-registry-wfwxw 1/1 Running 0 7m53s redhat-marketplace-94fmf 1/1 Running 0 36h redhat-marketplace-trgfg 1/1 Running 0 7m53s redhat-operators-t6bk5 1/1 Running 0 7m53s redhat-operators-wg6h9 1/1 Running 0 36h 3, According to the Comment 2 and Comment 7, none of the newer catalogs have an age of more than 15 minutes. And this is the original feature design. The bug is verified.