Environment degrade on image update

[lnetrebskii]

Steps to reproduce:

1. Update image name or tag in the kustomization "/platform-app/overlays/dev/kustomization.yaml"
2. Wait for synchronization

Actual result:
Error appears
FailedAttachVolume Multi-Attach error for volume "pvc-9816f174-2741-4569-9f1f-f922441bb720" Volume is already used by pod(s) platform-platform-55656b86ff-8t9xr

Expected result:
Container with the new image is started and the previous one has dropped

[Woland2k]

This is due to the usage of azure-disk as Persistent Volume. It only supports one mount at a time and since other POD Is still running it can't mount it. The reason I used azure-disk is that it is faster/cheaper than azurefiles (used for cms modules). Ideally we should switch to node based storage providers, for instance https://github.com/cybozu-go/topolvm or something similar.