[rabbitmq-discuss] Delay between minority detected and stopped server

[Malte Schirmacher]

On 14/08/13 22:50, Matthias Radestock wrote:
>> Suppose the publisher is running on the same machine as the broker and
>> someone pulls the network-cable out of this box.
>> Then it is pushed into minority instantaneously but would possibly still
>> accept messages via 127.0.0.1 leading to inconsistency between the
>> mirrored queues what we tried to prevent by using pause_minority...
>
> There is *always* scope for that, since publishing is inherently
> asynchronous, and so is detecting of cluster partitioning.

Yet there *must not* be scope for this. There is more then a minute 
delay between the node noticing it's in minority and actually shutting 
down the node. Furthermore there is nothing here to be explained simply 
by saying message delivery is asynchonous as delivering messages was 
startet AFTER the node logged that its in minority.
--
Geschaeftsanschrift/Business Address: crosscan GmbH | Ruhrstraße 48 | 58452 Witten | Germany
Support: +49.2302.28232-22 Phone: +49.2302.28232-00 Fax: +49.2302.28232-09 
Geschaeftsfuehrung/Management Board: Philip Lehmann, Erwin Berg, Ulrich Kellner
Sitz Witten, Amtsgericht Bochum, HRB 8036/Registered Office Witten, Commercial Register of the Bochum County Court, HRB 8036
UST-ID-Nr./VAT-IdNo.: DE234398770