Description
On our greenboard environment, when I was trying to deploy a function, the deployment is struck in deploying state for really long time without any deployment being actually started. No items for the functions are created yet in the metadata bucket. The same function code is working as expected when a second function is deployed.
Steps that led to this state:
1. While developing the eventing function, I had deployed and undeployed multiple times. All the deployment and undeployment went fine without any issue.
2. During the last undeployment and redeployment with a script change, the deployment is struck in deploying state for more than 12 hours.
3. I restarted eventing process on the machine.
4. Created a second function with the same script. The deployment goes through and working as expected.
The environment is still available in the same state for investigation. I will keep this for as long as required since other functions seem to working fine for now.
One other thing to note is that this deployment has more than 10 buckets in the cluster.
Also, the dcb backlog for the function was huge at about 60M+