Controller in non-default rack become unreachable via mcollective after netconfig task

[Vladimir Khlyunev]

ISO 1566

Steps to repro:
- Create env with at least 2 racks
- Add 1+ nodes with controller role to non-default rack
- Deploy cluster

Result:
After netconfig task node with non-default rack loses route to admin network in default rack whick leads to broken mcollective agent on this node:

E, [2017-04-06T16:43:41.348111 #7582] ERROR -- : rabbitmq.rb:30:in `on_miscerr' Unexpected error on connection stomp://mcollective@10.109.0.2:61613: es1_oldrecv: receive failed: Connection timed out @ io_fillbuf - fd:8
I, [2017-04-06T16:43:41.348364 #7582] INFO -- : rabbitmq.rb:10:in `on_connecting' TCP Connection attempt 0 to stomp://mcollective@10.109.0.2:61613
I, [2017-04-06T16:43:41.357711 #7582] INFO -- : rabbitmq.rb:25:in `on_connectfail' TCP Connection to stomp://mcollective@10.109.0.2:61613 failed on attempt 0
I, [2017-04-06T16:43:41.368031 #7582] INFO -- : rabbitmq.rb:10:in `on_connecting' TCP Connection attempt 1 to stomp://mcollective@10.109.0.2:61613
I, [2017-04-06T16:43:42.368709 #7582] INFO -- : rabbitmq.rb:25:in `on_connectfail' TCP Connection to stomp://mcollective@10.109.0.2:61613 failed on attempt 1
I, [2017-04-06T16:43:42.389123 #7582] INFO -- : rabbitmq.rb:10:in `on_connecting' TCP Connection attempt 2 to stomp://mcollective@10.109.0.2:61613

Actual IP network for this non-default rack - 10.109.5.0/24 and it reachable from this node; also manual update of mcollective config file fixes this:

D, [2017-04-06T17:03:39.332984 #9726] DEBUG -- : rabbitmq.rb:64:in `on_hbfire' Received heartbeat from stomp://mcollective@10.109.7.2:61613: receive_fire, {:curt=>1491498219.3327124}

Snapshot is not available right now because the env was passed to dev team

[Oleksiy Molchanov (omolchanov)]

No diagnostic snapshot, marking as Incomplete.