Only register the first web-dispatch listener

[lhoguin]

The problem: on certain double IPv4 and IPv6 environments, web-dispatch
will start a single listener (for management for example) but will
register two listeners in the Rabbit Mnesia table. Later, RabbitMQ
will crash when trying to stop or suspend them (rabbitmq-upgrade drain)
because only one of the two listeners exist in Ranch.

The fix: we register only the first listener, because that's the one
we used to create the Ranch ref (see rabbit_networking:ranch_ref/1).

Types of Changes

What types of changes does your code introduce to this project?
Put an x in the boxes that apply

• Bug fix (non-breaking change which fixes issue #NNNN)
• New feature (non-breaking change which adds functionality)
• Breaking change (fix or feature that would cause an observable behavior change in existing systems)
• Documentation improvements (corrections, new content, etc)
• Cosmetic change (whitespace, formatting, etc)

Checklist

Put an x in the boxes that apply. You can also fill these out after creating
the PR. If you're unsure about any of them, don't hesitate to ask on the
mailing list. We're here to help! This is simply a reminder of what we are
going to look for before merging your code.

• I have read the CONTRIBUTING.md document
• I have signed the CA (see https://cla.pivotal.io/sign/rabbitmq)
• All tests pass locally with my changes
• I have added tests that prove my fix is effective or that my feature works
• I have added necessary documentation (if appropriate)
• Any dependent changes have been merged and published in related repositories

Further Comments

Present in at least 3.8.11, tested on Windows.

@dumbbell is all over the git blame for the original implementation so a review would be great. Also I am not sure this is the right fix, I am wondering if it worked before and if perhaps something else broke it. Or perhaps the drain command did not exist at the time.

RabbitMQ and Ranch disagreeing on what listeners are available:

PS C:\Windows\system32> rabbitmqctl.bat eval 'rabbit_networking:node_listeners(node()).'
[{listener,rabbit@WinDev2104Eval,http,"WinDev2104Eval",
           {0,0,0,0,0,0,0,0},
           15672,
           [{cowboy_opts,[{sendfile,false}]},{port,15672}]},
 {listener,rabbit@WinDev2104Eval,http,"WinDev2104Eval",
           {0,0,0,0},
           15672,
           [{cowboy_opts,[{sendfile,false}]},{port,15672}]},
 {listener,rabbit@WinDev2104Eval,clustering,"WinDev2104Eval",
           {0,0,0,0,0,0,0,0},
           25672,[]},
 {listener,rabbit@WinDev2104Eval,amqp,"WinDev2104Eval",
           {0,0,0,0,0,0,0,0},
           5672,
           [{backlog,128},
            {nodelay,true},
            {linger,{true,0}},
            {exit_on_close,false}]},
 {listener,rabbit@WinDev2104Eval,amqp,"WinDev2104Eval",
           {0,0,0,0},
           5672,
           [{backlog,128},
            {nodelay,true},
            {linger,{true,0}},
            {exit_on_close,false}]}]
PS C:\Windows\system32> rabbitmqctl.bat eval 'ets:tab2list(ranch_server).'
[{{addr,{acceptor,{0,0,0,0},5672}},{{0,0,0,0},5672}},
 {{addr,{acceptor,{0,0,0,0,0,0,0,0},5672}},{{0,0,0,0,0,0,0,0},5672}},
 {{addr,{acceptor,{0,0,0,0,0,0,0,0},15672}},{{0,0,0,0},15672}},
 {{conns_sup,{acceptor,{0,0,0,0},5672}},<11164.749.0>},
 {{conns_sup,{acceptor,{0,0,0,0,0,0,0,0},5672}},<11164.734.0>},
 {{conns_sup,{acceptor,{0,0,0,0,0,0,0,0},15672}},<11164.621.0>},
...

25672 is the distribution so it's expected that it's not in Ranch. 5672 has two in both cases. 15672 has one in Ranch and two in RabbitMQ: this is the problem.

[lhoguin]

The same issue exists for the Prometheus listener. I will need to double check whether this PR fixes it as well.

[lhoguin]

I have now tested that this fixes the original issue. Please review.

[michaelklishin]

Backported to v3.8.x.