[rabbitmq-discuss] Hang while waiting for shutdown of last node in an HA cluster

[Matthias Radestock]

Matt,

On 16/07/12 21:23, Matt Pietrek wrote:
> We have a 3 node cluster running 2.8.4 supporting a small number of HA
> queues. During a controlled shutdown, we bring down each broker (play,
> play2, util) in sequence using something along the lines of
>
> rabbitmqctl stop <pidfile>
>
> The first two brokers shut down without incident. The last shutdown gets
> stuck and never finishes.
> [...]
> =ERROR REPORT==== 16-Jul-2012::13:06:19 ===
> ** Generic server rabbit_node_monitor terminating
> ** Last message in was {'DOWN',#Ref<0.0.0.1578>,process,
>                                 {rabbit,rabbit at play2},
>                                 normal}
> ** When Server state == [rabbit at play2]
> ** Reason for termination ==
> ** {bad_return_value,
>         {error,
>             {badarg,
>                 [{erlang,is_process_alive,[<3173.371.0>]},
>                  {rabbit_amqqueue,'-on_node_down/1-fun-1-',8},

Eeek. That's a bug that got introduced in 2.8.3. I cannot see atm why 
that would cause 'stop' to get stuck, but I wouldn't rule it out. Thanks 
for reporting this.

Regards,

Matthias.