Hide Forgot
Created attachment 1485266 [details] loglevel 8 migrate with errors Description of problem: During storage oc adm migrate storage, the tool is repeatedly reporting messages like: E0920 18:39:49.045977 error: -n hasha-pro1 pods/nodejs-ex-2-kmzrv: the server does not allow this method on the requested resource Version-Release number of selected component (if applicable): v3.11.7 How reproducible: 100% Additional info: High log level output attached.
https://github.com/openshift/origin/pull/21071
> openshift_metrics_server_install: "{{ openshift_metrics_install_metrics | default(False) }}" Removing this as a release-3.11 blocker since it only occurs if you opt-in to installing the metrics server.
Commit pushed to master at https://github.com/openshift/origin https://github.com/openshift/origin/commit/318343c755f9e47679c634216e2f25c8da4e8be3 migrate: ignore resources that cannot be listed and updated This change updates FindAllCanonicalResources to check the verbs asserted by discovery for listing and updating capability. This prevents us from trying to update aggregated resources such as pod metrics. Bug 1631517 Signed-off-by: Monis Khan <mkhan>
https://github.com/openshift/origin/pull/21099 should fix the terrible error messages as well.
Commit pushed to master at https://github.com/openshift/origin https://github.com/openshift/origin/commit/08e03204f63ce1d0dc2f23b21a5f5f05d41fb266 Use group resource in migrate reporting This change updates the reporting during migration to use the group resource instead of just the resource. Thus the confusing messages with "the server does not allow this method on the requested resource" for pod metrics will go from: -n hasha-pro1 pods/nodejs-ex-2-kmzrv to: -n hasha-pro1 pods.metrics.k8s.io/nodejs-ex-2-kmzrv This makes it clear what resource is being referenced. Bug 1631517 Signed-off-by: Monis Khan <mkhan>
Checked the latest build v3.11.16. It includes the 1st PR #21071's code but not the 2nd PR yet. A confusion: saw target release is 3.11.z but advisory/36292 attaches this bug. Pls either include the 2nd PR into 3.11 and update 3.11.z to 3.11.0 for verifying it in 3.11, or drop the bug from the advisory. Thanks
(In reply to Xingxing Xia from comment #7) > Checked the latest build v3.11.16. It includes the 1st PR #21071's code but > not the 2nd PR yet. A confusion: saw target release is 3.11.z but > advisory/36292 attaches this bug. Pls either include the 2nd PR into 3.11 > and update 3.11.z to 3.11.0 for verifying it in 3.11, or drop the bug from > the advisory. Thanks 2nd PR is just to improve error messaging in the future. It will not be included in 3.11 or 3.11.z. The 1st PR is the actual fix for this issue. It is not a blocker for 3.11, hence the 3.11.z. I have removed it from the advisory.
Now that the 1st PR goes into 3.11, VERIFIED the 1st PR with below steps. Given env is installed with: openshift_ansible_vars: openshift_metrics_server_install: true (or if the env is not installed with that, add these vars in inventory file, then run `ansible-playbook -i hosts /usr/share/ansible/openshift-ansible/playbooks/metrics-server/config.yml`) # oc get apiservice | grep metrics v1beta1.metrics.k8s.io 2018-09-29T06:31:55Z Use old oc 3.11.12 to run: $ oc adm migrate storage --loglevel=8 |& tee migrate.log $ grep "the server does not allow this" migrate.log There are many error logs. In another env, use latest oc 3.11.17 to run above, no such error logs. So the 1st PR is VERIFIED.
Checked with the latest 'openshift v3.11.36', the PR still not merged with ocp3.11.x.
Confirmed with openshift v4.0.0-0.49.0, the issue has fixed: [root@qe-yinzhou-master-etcd-1 ~]# openshift version openshift v4.0.0-0.49.0 [root@qe-yinzhou-master-etcd-1 ~]# oc adm migrate storage --loglevel=8 |& tee migrate.log [root@qe-yinzhou-master-etcd-1 ~]# cat migrate.log |grep unchanged |grep images I1107 02:43:28.376186 unchanged: -n zhouyt imagestreams.image.openshift.io/ruby-25-centos7 I1107 02:43:28.376496 unchanged: -n zhouy imagestreams.image.openshift.io/mytest
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