Host in maintenance mode in zone 2 enters alert state due to Persistent network present on zone 1

[Pearl1594]

problem

There seems to be an issue with the way Persistent networks get setup.

versions

ACS 4.22 (but maybe earlier too)

The steps to reproduce the bug

1. Put the host in Zone 2 into maintenance.
2. Add L2 network with a vlan tag in Zone 1.
3. While the host is UP and in Maintenance state, restart the cloudstack-agent. It will transition into Alert state.
4. The cloudstack-agent will keep in a loop state trying to connect to Management server, if we delete the L2 network from Zone 1, cloudstack-agent connects to the Management server and transition into UP state.
5. Exit maintenance status manually and the host works fine even if we add the L2 network again and restart the cloudstack-agent.

What to do about it?

Persistent networks should adhere to zone isolation - those created in one zone shouldn't be attempted to be created on another.