[rabbitmq-discuss] RabbitMQ Clustering in Windows Server

Emile Joubert emile at rabbitmq.com
Wed Aug 25 10:01:14 BST 2010


Hi Steve,

There may have been some confusion between rabbit clustering and
Microsoft Cluster Service - they are completely different. Rabbit
clustering is based on the Erlang distributed database and is fully
supported and widely used on Windows.

MSCS can be used to set up high availability services, while rabbit
clustering is used to achieve scalability. It should be possible to set
up MSCS in single quorum mode using shared storage to get HA in the same
way as Pacemaker, but I'm not aware of anyone trying this. We would be
interested to hear about your experience if you try this or succeed.

Regards

Emile




On 24/08/10 20:45, Steve Follett wrote:
> David,  the question Desrie poses is mine as well.  Those of us that have done Microsoft clustering for some time understand there are a few simple things needed to allow an application to be configured on MS cluster, and most seem to be readily available in RabbitMQ windows version.  We are also looking into the Linux Pacemaker configuration as well but thought we could hit two paths for completeness.
>
>
> For Microsoft clustering an application it needs to be able to run as a service.
> It needs to be able to listen on a unique configured IP address rather than the local host, which seems to be able to be configured in RabbitMQ.
> It needs to be able to configure all relevant log files onto the shared storage location, which seems to be able to be configured in RabbitMQ.
>
> With these as a basis do you know of anyone who has tried this kind of configuration rather than the Windows version posted on your site?
>
>
> Steve
>
> Hi Desrie,
>
>
> Desrie Qi <desrie.qi at wealthcraft.com<https://lists.rabbitmq.com/cgi-bin/mailman/listinfo/rabbitmq-discuss>> writes:
>
>   
>> We run RabbitMQ as windows service. Does anyone cluster RabbitMQ service in Windows server?
>>     
>> Should we cluster two RabbitMQ nodes following two steps bellow or either one is OK?
>>     
>> Step 1) Join two nodes in a cluster by RabbitMQ command: rabbitmqctl
>>     
>>       @rabbit2\\C:\RabbitMQ_server-1.8.1\sbin> rabbitmqctl cluster rabbit at rabbit1<https://lists.rabbitmq.com/cgi-bin/mailman/listinfo/rabbitmq-discuss>
>>     
>>       Clustering node rabbit at rabbit2<https://lists.rabbitmq.com/cgi-bin/mailman/listinfo/rabbitmq-discuss> with [rabbit at rabbit1<https://lists.rabbitmq.com/cgi-bin/mailman/listinfo/rabbitmq-discuss>] ...done.
>> Step 2) cluster two RabbitMQ service using Windows clustering manager.
>>     
> I'm not sure what the "Windows clustering manager" you refer to is.  But
> the instructions for clustering at
> <http://www.rabbitmq.com/clustering.html> should apply on Windows,
> including in the case where you run the nodes as Windows services.  Is
> there a specific problem you are having when trying to get this working?
>
>
> David
>
>
> --
>
> David Wragg
>
> Staff Engineer, RabbitMQ
>
> SpringSource, a division of VMware
>
>
>
> Steve Follett
> Director Technical Operations | www.Troppussoftware.com<http://www.troppussoftware.com/>
> Mobile: (303)803-7812 | steve.follett at troppussoftware.com<mailto:stevefollett at peak8solutions.com>



More information about the rabbitmq-discuss mailing list