Red Hat Bugzilla – Bug 1559271
AWS nodes does not keep persistent labels after shut down
Last modified: 2018-10-11 03:20:12 EDT
Description of problem: OCP nodes on AWS do not keep persistent labels when given from "oc label" command after shutting down the node. Version-Release number of selected component (if applicable): oc v3.6.173.0.21 kubernetes v1.6.1+5115d708d7 Actual results: Labels are not persistent after node shutdown using "oc command" Expected results: Should be persistent Additional Info : Correct entries in /etc/origin/cloudprovider/aws.conf Same Kubernetes Cluster Tags across all cluster
Recently had a similiar issue with Openstack cloudprovider in that the node controller was getting a filtered list of instances, only getting the active ones. https://bugzilla.redhat.com/show_bug.cgi?id=1558422
More relevant upstream discussion/work: https://github.com/kubernetes/kubernetes/issues/58635 https://github.com/kubernetes/kubernetes/pull/60009 https://github.com/kubernetes/kubernetes/issues/46442 https://github.com/kubernetes/kubernetes/issues/45118 There is also an old open PR to remove the instance filtering from the AWS cloudprovider, which would address is bug: https://github.com/kubernetes/kubernetes/pull/45986
The effort has stalled upstream. I am pushing for acceptance of a targeted fix for just this case (one line change in the AWS cloudprovider) as a stopgap while a more general solution is developed.
*** Bug 1569043 has been marked as a duplicate of this bug. ***
*** Bug 1608971 has been marked as a duplicate of this bug. ***
*** Bug 1608912 has been marked as a duplicate of this bug. ***
I posted a more concise PR upstream to avoid friction: https://github.com/kubernetes/kubernetes/pull/66835
Origin PR: https://github.com/openshift/origin/pull/20663
*** Bug 1619599 has been marked as a duplicate of this bug. ***
Checked on # oc version oc v3.11.0-0.28.0 kubernetes v1.11.0+d4cacc0 features: Basic-Auth GSSAPI Kerberos SPNEGO Server https://ip-172-18-7-85.ec2.internal:8443 openshift v3.11.0-0.28.0 kubernetes v1.11.0+d4cacc0 And this issue has been fixed. 1. labels from `oc label` is persistent now. 2. labels from node-config.yaml(from 3.10, should be configmap in openshift-node namespace) is also applied after reboot system 3. labels already applied to node from node-config.yaml will still keep after reboot the system
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-2018:2652