thanks so much for any guidance on this.
I have a deployment with 1 replica. I'm in a dev lab environment with a brand new Tanzu k8s cluster. Here are the steps I've taken and a curious problem I've noticed.
SQL server 2019 is successfully in running state and can be connected with Azure data studio, data can be read and written no problem.
To simulate catastrophic failue, the SQl server pod is deleted from the command line while Azure data studio is connected
This results in a dangling volume that is eventually mounted by the new pod on a new node but takes upwards of 4 to 5 minutes to return to running state and be available again. After the pod is in running state, there is no problem, the weather is fine.
mssql-deployment-8bc6c96dc-rqxlj 0/1 Pending 0 0s <none> <none> <none> <none>
mssql-deployment-8bc6c96dc-rqxlj 0/1 Pending 0 0s <none> sqltkc-workers-mpqdb-556696d6f6-rvs26 <none> <none>
mssql-deployment-8bc6c96dc-rqxlj 0/1 ContainerCreating 0 0s <none> sqltkc-workers-mpqdb-556696d6f6-rvs26 <none> <none>
mssql-deployment-8bc6c96dc-rqxlj 1/1 Running 0 4m11s 192.168.1.14 sqltkc-workers-mpqdb-556696d6f6-rvs26 <none> <none>
My question is:
1) Why does it take so long?
2) Is there a way to speed this up through a config setting in the deployment or PVC?
3) Is there a way to fully terminate a pod BEFORE you spin up a new one and attempt to mount the volume?
Thanks!!!