+++ This bug was initially created as a clone of Bug #2094460 +++
Description of problem:
When addNode() failed in addNodeAnnotations(), the node's IPAM can be
overwritten by subsequent addNode() retry attempts. As the result, the
same IP can be allocated to multiple pods.
retryPod worker is started too soon, which could start to handle add Pod
request and allocate duplicate Pod IPs before IPs of the existing Pods are
reserved in syncPods().
Version-Release number of selected component (if applicable):
Steps to Reproduce:
No duplicate IPs
Fixed in upstream PR: https://github.com/ovn-org/ovn-kubernetes/pull/2928
We may also need
--- Additional comment from Martin Kennelly on 2022-06-07 16:22:14 UTC ---
Fix merged downstream. Pending release of 4.11.
Can you put 'cherry-pick-approved' label on the PR?
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 (OpenShift Container Platform 4.10.20 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.