sidshas03 opened a new pull request, #14701:
URL: https://github.com/apache/cloudstack/pull/14701

   upgrade-kubernetes.sh imports container images only on the node it runs on.
   During CKS upgrade the control node gets cordoned first, kubeadm creates
   a health-check Job with the new pause image, Job gets scheduled on a worker
   that still has the old pause image, pull fails because no internet 
(air-gapped),
   upgrade fails after all retries.
   
   This adds an IMPORT_IMAGES_ONLY mode to the upgrade script and a pre-staging
   step in KubernetesClusterUpgradeWorker that runs the script on all workers
   after ISO attach but before the control node upgrade starts. Also switched
   PAUSE_IMAGE detection from ctr images ls to kubeadm config images list and
   started logging script output on failures which was getting swallowed 
earlier.
   
   Fixes: #14654
   - [x] Bug fix (non-breaking change which fixes an issue)
   - [x] Major
   
   Tested on offline CKS cluster upgrading from v1.35.9 to v1.36.5 with
   1 control + 2 workers, Cilium CNI, containerd 1.7.12, Ubuntu 24.04.
   Verified workers had pause:3.10.2 imported before control node ran
   kubeadm upgrade apply. Health-check Job completed without image pull
   errors. Checked management-server.log shows script output on simulated
   failures.


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]

Reply via email to