SQL pod may stuck in « ContainerCreating » status when you stop the node instance on AKS


When we deploy SQL Server on AKS, sometimes we may find SQL HA is not working as expect.

 

For example, when we deploy AKS using our default sample with 2 nodes:

https://docs.microsoft.com/en-us/azure/aks/tutorial-kubernetes-deploy-cluster#create-a-kubernetes-cluster

 

 

az aks create     –resource-group myResourceGroup     –name myAKSCluster     –node-count 2     –generate-ssh-keys     –attach-acr <acrName>

 

 

There should be 2 instances deployed in the AKS virtual machine scale set:

 

According to the SQL document:

 

In the following diagram, the node hosting the mssql-server container has failed. The orchestrator starts the new pod on a different node, and mssql-server reconnects