doesn't delete old node's IQN initiator

[inajob]

Describe the bug

I rolling updated my kubernetes cluster. My updatig way is createing new Node and joining the cluster and deleteing old Node again and again.

I also use trident and volume backed by ontap-san.
I noticed that my igroup has old node's IQN initiators yet.

Environment

• Trident version: 20.01.1
• Trident installation flags used: trident install -n trident
• Container runtime: -
• Kubernetes version: 1.17.2
• Kubernetes orchestrator: original
• Kubernetes enabled feature gates: -
• OS: ContainerLinux
• NetApp backend types: -
• Other:

To Reproduce

As I mentioned, add new Node and drain and delete old Node.

Expected behavior

igroup has only IQN initiators which is a member of the cluster.

Additional context

We automate this rolling update method, so we have to delete old IQN initiator manually.
I want trident to delete old ones.

[adkerr]

This is actually working as designed, as when the code was implemented we didn't have a great way to reliably detect if/when it was safe to remove the iqn. That said, we do plan to implement the auto-removal of inactive iqns in a future update.

[inajob]

@adkerr
OK! I'm looking forward to implement the auto-removal feature!

[inajob]

@adkerr
When will you implement this feature?
I would like to use this feature in next release(20.04.0).

[gnarl]

This enhancement is included in the Trident 20.07 release.