Description of problem: Operators that can be installed on 4.3 are not installing on 4.1.z Version-Release number of selected component (if applicable): 4.1.22 How reproducible: Can reproduce on specific operators. Steps to Reproduce: 1. Install an OpenShift 4.1.22 cluster 2. Attempt to install the service mesh operator provided by the Redhat-Operators OperatorSource. Actual results: The subscription is created but the operator a CSV is not created - as a result no operator is deployed. Expected results: The service mesh operator is successfully installed. Additional info: Originally created when investigating https://bugzilla.redhat.com/show_bug.cgi?id=1763750#c2
The servicemesh operator depends on a number of operator, and the catalog containing those operators need to be present in the namespace that servicemesh is being installed in. To install the servicemesh operator 1) Click install in the UI 2) Copy over the catalogsource to the openshift-operators namespace 3) Check for running servicemesh operator pods in the openshift-operators namespace As such, this is working as intended in 4.1.z clusters.
Hi, Nick Thanks for your explanation! > When using the console in 4.1, any Subscription to an operator that has any dependencies will not be installed, this is expected behavior... So, that means we don't support the ServiceMesh operator in 4.1, right? I have a fresh 4.1.z cluster. And, the ServiceMesh doesn't work. Cluster version is 4.1.0-0.nightly-2019-12-04-071458 mac:~ jianzhang$ oc get sub -n openshift-operators NAME PACKAGE SOURCE CHANNEL servicemeshoperator servicemeshoperator installed-redhat-openshift-operators 1.0 mac:~ jianzhang$ oc get csv -n openshift-operators No resources found in openshift-operators namespace. mac:~ jianzhang$ oc get csc -n openshift-marketplace NAME STATUS MESSAGE AGE certified-operators Succeeded The object has been successfully reconciled 56m community-operators Succeeded The object has been successfully reconciled 56m installed-redhat-openshift-operators Succeeded The object has been successfully reconciled 15m redhat-operators Succeeded The object has been successfully reconciled 56m mac:~ jianzhang$ oc get catalogsource -n openshift-operators NAME NAME TYPE PUBLISHER AGE installed-redhat-openshift-operators Red Hat Operators grpc Red Hat 15m Catalog operator logs: mac:~ jianzhang$ oc logs catalog-operator-5b79fb7c46-2gr9r ... time="2019-12-05T08:29:52Z" level=info msg="retrying openshift-operators" E1205 08:29:52.952693 1 queueinformer_operator.go:186] Sync "openshift-operators" failed: {servicemeshoperator 1.0 servicemeshoperator.v1.0.2 {installed-redhat-openshift-operators openshift-operators}} not found: rpc error: code = Unavailable desc = all SubConns are in TransientFailure, latest connection error: connection error: desc = "transport: Error while dialing dial tcp 172.30.175.186:50051: connect: no route to host" @Nick, In the above logs, why here are the `openshift-operators`? I think it should be `openshift-marketplace`, right? mac:~ jianzhang$ oc get svc -n openshift-marketplace NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE certified-operators ClusterIP 172.30.230.118 <none> 50051/TCP 48m community-operators ClusterIP 172.30.189.98 <none> 50051/TCP 48m installed-redhat-openshift-operators ClusterIP 172.30.175.186 <none> 50051/TCP 7m24s redhat-operators ClusterIP 172.30.130.232 <none> 50051/TCP 48m mac:~ jianzhang$ oc get pods -n openshift-marketplace -o wide NAME READY STATUS RESTARTS AGE IP NODE NOMINATED NODE READINESS GATES certified-operators-597f9c4c74-bvgdl 1/1 Running 0 44m 10.128.2.6 ip-10-0-142-202.ec2.internal <none> <none> community-operators-67849c6d56-47r5x 1/1 Running 0 44m 10.128.2.7 ip-10-0-142-202.ec2.internal <none> <none> installed-redhat-openshift-operators-768b56448-9bm8r 1/1 Running 0 3m25s 10.129.2.8 ip-10-0-136-31.ec2.internal <none> <none> marketplace-operator-849b56d765-hrjg8 1/1 Running 0 44m 10.128.0.15 ip-10-0-147-113.ec2.internal <none> <none> redhat-operators-785b67d7c4-gc7vg 1/1 Running 0 44m 10.128.2.3 ip-10-0-142-202.ec2.internal <none> <none> mac:~ jianzhang$ oc -n openshift-sdn ovs-87v55 Error: unknown command "ovs-87v55" for "oc" Run 'oc --help' for usage. mac:~ jianzhang$ oc -n openshift-sdn rsh ovs-87v55 sh-4.2# ping 10.129.2.8 PING 10.129.2.8 (10.129.2.8) 56(84) bytes of data. 64 bytes from 10.129.2.8: icmp_seq=1 ttl=64 time=0.905 ms 64 bytes from 10.129.2.8: icmp_seq=2 ttl=64 time=0.208 ms 64 bytes from 10.129.2.8: icmp_seq=3 ttl=64 time=0.214 ms 64 bytes from 10.129.2.8: icmp_seq=4 ttl=64 time=0.222 ms 64 bytes from 10.129.2.8: icmp_seq=5 ttl=64 time=0.211 ms mac:~ jianzhang$ oc rsh catalog-operator-5b79fb7c46-2gr9r sh-4.2$ nslookup 172.30.175.186 186.175.30.172.in-addr.arpa name = installed-redhat-openshift-operators.openshift-marketplace.svc.cluster.local. That `installed-redhat-openshift-operators-768b56448-9bm8r` pod and DNS works well, why get the `Error while dialing dial tcp 172.30.175.186:50051: connect: no route to host`?