Hide Forgot
Description of problem: After the heat stack was created, S2I build was failed due to the docker-registry IP had been changed. And the changes was done by extra playbook, not openshift-ansible. See https://bugzilla.redhat.com/show_bug.cgi?id=1192257 Version-Release number of selected component (if applicable): v0.9.2 How reproducible: Sometimes Steps to Reproduce: 1.Create a heat stack 2.Do S2I build after the stack was completed 3. Actual results: 1. docker-registry and router were created successfully, but found that docker-registry was re-deployed. # oc get po -n default NAME READY STATUS RESTARTS AGE docker-registry-2-8pk96 1/1 Running 0 45m router-1-n8qva 1/1 Running 0 45m router-1-qnnnb 1/1 Running 0 44m # oc get svc -n default NAME CLUSTER-IP EXTERNAL-IP PORT(S) AGE docker-registry 172.30.138.108 <none> 5000/TCP 50m kubernetes 172.30.0.1 <none> 443/TCP,53/UDP,53/TCP 1h router 172.30.29.192 <none> 80/TCP,443/TCP,1936/TCP 50m 2. S2I build failed due to docker registry ip changed. I1017 03:14:08.420080 1 sti.go:334] Successfully built test/cakephp-example-1:4ddc039c I1017 03:14:08.662623 1 cleanup.go:23] Removing temporary directory /tmp/s2i-build260843205 I1017 03:14:08.662694 1 fs.go:156] Removing directory '/tmp/s2i-build260843205' I1017 03:14:08.721730 1 sti.go:268] Using provided push secret for pushing 172.30.138.108:5000/test/cakephp-example:latest image I1017 03:14:08.721752 1 sti.go:272] Pushing 172.30.138.108:5000/test/cakephp-example:latest image ... I1017 03:14:15.908876 1 sti.go:277] Registry server Address: I1017 03:14:15.908922 1 sti.go:278] Registry server User Name: serviceaccount I1017 03:14:15.908929 1 sti.go:279] Registry server Email: serviceaccount I1017 03:14:15.908935 1 sti.go:284] Registry server Password: <<non-empty>> F1017 03:14:15.908943 1 builder.go:204] Error: build error: Failed to push image. Response from registry is: Error: Status 404 trying to push repository test/cakephp-example: "404 page not found\n" Expected results: S2I build successed after the stack was created. Additional info: Seen from the ansible logs The docker-registry ip might be changed in this situation: After running BYO playbook (openshift-ansible), the task "Restart node service" was exected in extra playbook, this task *might* destroy the docker-registry deployment, then trigger a new deployment for docker-registry.
Created attachment 1211239 [details] Part logs of ansible
Close it as it can't be reproduced any more.