VMware Tanzu Kubernetes Grid Integrated Edition (TKGI) upgrade-cluster errand fails due to timeout when stopping dockerd

Products

• VMware Tanzu Kubernetes Grid Integrated Edition

Issue/Introduction

When running the "pks upgrade-cluster" errand, it hits an error trying to stop the dockerd process due to a 60 second timeout.

You can workaround this issue by doing "monit unmonitor all" and "monit stop all" on all of the Worker nodes and then running "pks upgrade-cluster" again.

Resolution

Workaround

The current workaround for this issue is to ssh into the affected nodes and stopping the monit processes.
 

Steps

1. Identify the node having the issue.
 

2. Run "bosh ssh" to ssh into the node.
 

3. Run the command: sudo su -
 

4. Run the command: monit unmonitor all
 

5. Run the command: monit stop all

6. Exit out of the node and re-run the command: pks upgrade-cluster