<div>The clustering + mirrored queue approach you've already got is what is required to make RabbitMQ HA, you don't (additionally) need the Heartbeat/DRBD�shenanigans, get rid of that as it just complicates things.</div>
<div><br></div><div>Since you also mention making this transparent to the clients (which is a�separate�issue), you will need to either embed the failover logic in the client code (making clients aware of every node in the cluster), or use a VIP, Proxy or something similar to present a single address/endpoint to the clients.</div>
<div><br></div><div>Many people have success using a TCP load-balancing mechanism like LVS or HAProxy in front of RabbitMQ.</div><div><br></div><div>This way if one of the nodes in your cluster goes down, the clients TCP connection to the dead node is reset and when they reconnect to the Proxy they will be directed to the healthy node, upon which the mirrored queue would be master.</div>
<div><br></div><div><br></div><div><div class="gmail_quote">On Thu, May 17, 2012 at 2:08 PM, sameekmishra <span dir="ltr"><<a href="mailto:sameek@arosys.com" target="_blank">sameek@arosys.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
<br>
I have implement the following points :<br>
<br>
1.clustering with two node<br>
2.mirror queue<br>
3. installation and configuration of heartbeat on both nodes<br>
4. installation of pacemaker.<br>
<br>
I need to implement: if any node is failed to switch to other node without<br>
know the client.<br>
<br>
as you say i have read it document-<a href="http://www.rabbitmq.com/pacemaker.html" target="_blank">http://www.rabbitmq.com/pacemaker.html</a><br>
but i can't able to find out without using DRBD.<br>
please help me how i configure it.<br>
<div class="im"><br>
<br>
<br>
Brendan Hay-2 wrote:<br>
><br>
> The document that you must have read which refers to DRBD is:<br>
> <a href="http://www.rabbitmq.com/pacemaker.html" target="_blank">http://www.rabbitmq.com/pacemaker.html</a><br>
><br>
</div>> Please read about M*irrored Queue Behaviour* via the link Simon mentioned<br>
<div class="im">> here: <a href="http://www.rabbitmq.com/ha.html" target="_blank">http://www.rabbitmq.com/ha.html</a><br>
><br>
> This will give you information about creating a highly available<br>
</div>> active/active setup *without* using something like DRBD.<br>
<div><div class="h5">><br>
> - Brendan<br>
><br>
> On Thu, May 17, 2012 at 12:56 PM, sameekmishra <<a href="mailto:sameek@arosys.com">sameek@arosys.com</a>> wrote:<br>
><br>
>><br>
>> I have already read this document.i want to know that DRBD required<br>
>> partition<br>
>> or it uses any free space in hard disk to stored file.if we can do then<br>
>> please explain it.otherwise tell me the approach to followed to achieve<br>
>> high<br>
>> availability without using DRBD.<br>
>><br>
>> Thanks<br>
>><br>
>><br>
>><br>
>> Simon MacMullen-2 wrote:<br>
>> ><br>
>> > On 17/05/12 05:27, sameekmishra wrote:<br>
>> >> can we manage the high availability without using DRBD,because have<br>
>> not<br>
>> >> created any shared partition for DRBD device.please help how can i<br>
>> >> implment<br>
>> >> high �availability.<br>
>> ><br>
>> > See <a href="http://www.rabbitmq.com/ha.html" target="_blank">http://www.rabbitmq.com/ha.html</a><br>
>> ><br>
>> > Cheers, Simon<br>
>> ><br>
>> > --<br>
>> > Simon MacMullen<br>
>> > RabbitMQ, VMware<br>
>> > _______________________________________________<br>
>> > rabbitmq-discuss mailing list<br>
>> > <a href="mailto:rabbitmq-discuss@lists.rabbitmq.com">rabbitmq-discuss@lists.rabbitmq.com</a><br>
>> > <a href="https://lists.rabbitmq.com/cgi-bin/mailman/listinfo/rabbitmq-discuss" target="_blank">https://lists.rabbitmq.com/cgi-bin/mailman/listinfo/rabbitmq-discuss</a><br>
>> ><br>
>> ><br>
>><br>
>> --<br>
>> View this message in context:<br>
>> <a href="http://old.nabble.com/can-we-implement-high-availability-without-using-DRBD--tp33862277p33863431.html" target="_blank">http://old.nabble.com/can-we-implement-high-availability-without-using-DRBD--tp33862277p33863431.html</a><br>
>> Sent from the RabbitMQ mailing list archive at Nabble.com.<br>
>><br>
>> _______________________________________________<br>
>> rabbitmq-discuss mailing list<br>
>> <a href="mailto:rabbitmq-discuss@lists.rabbitmq.com">rabbitmq-discuss@lists.rabbitmq.com</a><br>
>> <a href="https://lists.rabbitmq.com/cgi-bin/mailman/listinfo/rabbitmq-discuss" target="_blank">https://lists.rabbitmq.com/cgi-bin/mailman/listinfo/rabbitmq-discuss</a><br>
>><br>
><br>
> _______________________________________________<br>
> rabbitmq-discuss mailing list<br>
> <a href="mailto:rabbitmq-discuss@lists.rabbitmq.com">rabbitmq-discuss@lists.rabbitmq.com</a><br>
> <a href="https://lists.rabbitmq.com/cgi-bin/mailman/listinfo/rabbitmq-discuss" target="_blank">https://lists.rabbitmq.com/cgi-bin/mailman/listinfo/rabbitmq-discuss</a><br>
><br>
><br>
<br>
--<br>
</div></div>View this message in context: <a href="http://old.nabble.com/can-we-implement-high-availability-without-using-DRBD--tp33862277p33863776.html" target="_blank">http://old.nabble.com/can-we-implement-high-availability-without-using-DRBD--tp33862277p33863776.html</a><br>
<div class="HOEnZb"><div class="h5">Sent from the RabbitMQ mailing list archive at Nabble.com.<br>
<br>
_______________________________________________<br>
rabbitmq-discuss mailing list<br>
<a href="mailto:rabbitmq-discuss@lists.rabbitmq.com">rabbitmq-discuss@lists.rabbitmq.com</a><br>
<a href="https://lists.rabbitmq.com/cgi-bin/mailman/listinfo/rabbitmq-discuss" target="_blank">https://lists.rabbitmq.com/cgi-bin/mailman/listinfo/rabbitmq-discuss</a><br>
</div></div></blockquote></div><br></div>