Description of problem: When configuring a bridge device using ip (netlink) tool - nmastate-handler on the node enters CrahsLoopBackOff. Version-Release number of selected component (if applicable): OCP Version: 4.7.0-fc.4 Kubernetes Version: v1.20.0+f0a2ec9 CNV Vesion: 2.6.0 nmstate verion: nmstate-0.3.4-17.el8_3.noarch How reproducible: Always Steps to Reproduce: 1. In the cluster - login to one of the worker nodes. [cnv-qe-jenkins@network02-khphv-executor ~]$ oc get nodes -l node-role.kubernetes.io/worker NAME STATUS ROLES AGE VERSION network02-khphv-worker-0-4zmxs Ready worker 2d4h v1.20.0+d9c52cc network02-khphv-worker-0-7q8sr Ready worker 2d4h v1.20.0+d9c52cc network02-khphv-worker-0-cmqsb Ready worker 2d4h v1.20.0+d9c52cc [cnv-qe-jenkins@network02-khphv-executor ~]$ [cnv-qe-jenkins@network02-khphv-executor ~]$ oc debug node/network02-khphv-worker-0-cmqsb Starting pod/network02-khphv-worker-0-cmqsb-debug ... To use host binaries, run `chroot /host` Pod IP: 192.168.3.248 If you don't see a command prompt, try pressing enter. sh-4.4# chroot /host sh-4.4# 2. Add a bridge device using ip tool: sh-4.4# ip link add name br-test type bridge Actual results: nmstate-handler pod on the node enters CrashLoopBackOff state. [cnv-qe-jenkins@network02-khphv-executor ~]$ oc get pod -n openshift-cnv -l component=kubernetes-nmstate-handler -o wide NAME READY STATUS RESTARTS AGE IP NODE NOMINATED NODE READINESS GATES nmstate-handler-2s6ff 1/1 Running 0 2d3h 192.168.2.23 network02-khphv-master-2 <none> <none> nmstate-handler-5g4jj 1/1 Running 0 2d3h 192.168.0.132 network02-khphv-master-1 <none> <none> nmstate-handler-9c8hs 1/1 Running 1 2d3h 192.168.0.14 network02-khphv-worker-0-7q8sr <none> <none> nmstate-handler-ggfn7 1/1 Running 1 2d3h 192.168.2.46 network02-khphv-worker-0-4zmxs <none> <none> nmstate-handler-r2kcd 1/1 Running 0 2d3h 192.168.1.146 network02-khphv-master-0 <none> <none> nmstate-handler-vt2ck 0/1 CrashLoopBackOff 18 3h42m 192.168.3.248 network02-khphv-worker-0-cmqsb <none> <none> Additional info: 1. nmstate-handler log and journalctl (from the node) are attached. 2. Workaround - restart the damaged pod by deleting it: [cnv-qe-jenkins@network02-khphv-executor yossi]$ oc delete pod -n openshift-cnv nmstate-handler-vt2ck pod "nmstate-handler-vt2ck" deleted
Created attachment 1754342 [details] nmstate-handler.log
Created attachment 1754343 [details] journalctl.log
Verified on: CNV 2.6.0 kubernetes-nmstate-handler-container-v2.6.0-18 (sha256:ed5fc663a75c40c878da3ce85e85e2a0b27f4e8cf02f93d9822cfb627a8266b2). by following the same scenario as in the bug description. The nmstate-handler pod remained stable, without restarting or crashing.
*** Bug 1923978 has been marked as a duplicate of this bug. ***
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 (Moderate: OpenShift Virtualization 2.6.0 security and bug fix update), 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/RHSA-2021:0799