Failed to deploy SystemVM resources are not removed from Tungsten Fabric

[rajujith]

ISSUE TYPE

• Bug Report

COMPONENT NAME

Tungsten Fabric, SystemVMs

CLOUDSTACK VERSION

4.18.0

CONFIGURATION

CloudStack advanced zone with Tungsten Fabric integration.

OS / ENVIRONMENT

SUMMARY

If the systemVMs fails to get deployed due to any reason the resources such as IP addresses and corresponding objects created by CloudStack at the Tungsten Fabric are not deleted when the SystemVMs are deleted from CloudStack. This results in subsequent deployment failure with the error Failure message: IP address already in use.

STEPS TO REPRODUCE

1. Create a zone with Tungsten Fabric isolation. 
2. Skip enabling the Tungsten Fabric plugin in the global settings: `tungsten.plugin.enable`
3. SystemVMs fail to deploy.
4. Enable the plugin above and restart cloudstack-management service.
5. Delete the existing SystemVMs
6. New SystemVMs fail to deploy with the error `Failure message: IP address already in use`. 

EXPECTED RESULTS

CloudStack should remove the failed-to-deploy SystemVMs IP addresses from Tungsten Fabric.

ACTUAL RESULTS

CloudStack is not removing the failed-to-deploy SystemVMs IP addresses from Tungsten Fabric.

WORKAROUND

Disable the zone
Delete the existing SystemVMs

Login to the Tungsten controller and click on configure. 
Networking -> ports.

Select all and delete all.