<br><br>On Monday, November 7, 2011 4:06:53 AM UTC+5:30, Konstantin Kalin wrote:<blockquote class="gmail_quote" style="margin: 0;margin-left: 0.8ex;border-left: 1px #ccc solid;padding-left: 1ex;"><p>On Oct 2, 9:12 am, Matthew Sackman <<a>matt...@rabbitmq.com</a>> wrote:<br>> On Sat, Oct 01, 2011 at 09:18:17AM -0700, Alex Grönholm wrote:<br>> > I noticed that the hot new feature in RabbitMQ 2.6 is Active/Active HA on<br>> > the server side. How can I take advantage of this on the client side? AFAIK<br>> > the client only knows how to connect to a single server -- you can't give it<br>> > a list of servers for failover purposes.<br>><br>> Correct. The Active/Active HA work does not address the issue of client<br>> connectivity. There are a number of solutions that you could use:<br>><br>> o) Change the DNS on failure<br>> o) a plain TCP load balancer like haproxy or a router-based one<br>> o) mobile IP addresses as can be done with something like pacemaker</p><p>I think there is one more valuable option. If a client knows cluster's<br>nodes (IP addresses) the client can reconnect to another node. I<br>implemented this in ~150 lines in Java code when I decided to test<br>RabbitMQ Active/Active. And it works fine in my opinion</p><p>Thank you,<br>Konstantin.<br>______________________________<wbr>_________________<br>rabbitmq-discuss mailing list<br><a href="javascript:" target="_blank" gdf-obfuscated-mailto="GapEMEEBeRwJ">rabbitmq...@lists.<wbr>rabbitmq.com</a><br><a href="https://lists.rabbitmq.com/cgi-bin/mailman/listinfo/rabbitmq-discuss" target="_blank">https://lists.rabbitmq.com/<wbr>cgi-bin/mailman/listinfo/<wbr>rabbitmq-discuss</a><br></p><p></p><p></p></blockquote><br>On Monday, November 7, 2011 4:06:53 AM UTC+5:30, Konstantin Kalin wrote:<blockquote class="gmail_quote" style="margin: 0;margin-left: 0.8ex;border-left: 1px #ccc solid;padding-left: 1ex;"><p>On Oct 2, 9:12 am, Matthew Sackman <<a>matt...@rabbitmq.com</a>> wrote:<br>> On Sat, Oct 01, 2011 at 09:18:17AM -0700, Alex Grönholm wrote:<br>> > I noticed that the hot new feature in RabbitMQ 2.6 is Active/Active HA on<br>> > the server side. How can I take advantage of this on the client side? AFAIK<br>> > the client only knows how to connect to a single server -- you can't give it<br>> > a list of servers for failover purposes.<br>><br>> Correct. The Active/Active HA work does not address the issue of client<br>> connectivity. There are a number of solutions that you could use:<br>><br>> o) Change the DNS on failure<br>> o) a plain TCP load balancer like haproxy or a router-based one<br>> o) mobile IP addresses as can be done with something like pacemaker</p><p>I think there is one more valuable option. If a client knows cluster's<br>nodes (IP addresses) the client can reconnect to another node. I<br>implemented this in ~150 lines in Java code when I decided to test<br>RabbitMQ Active/Active. And it works fine in my opinion</p><p>Thank you,<br>Konstantin.<br>______________________________<wbr>_________________<br>rabbitmq-discuss mailing list<br><a href="javascript:" target="_blank" gdf-obfuscated-mailto="GapEMEEBeRwJ">rabbitmq...@lists.<wbr>rabbitmq.com</a><br><a href="https://lists.rabbitmq.com/cgi-bin/mailman/listinfo/rabbitmq-discuss" target="_blank">https://lists.rabbitmq.com/<wbr>cgi-bin/mailman/listinfo/<wbr>rabbitmq-discuss</a></p></blockquote><div><div><font face="Calibri">Hi Konstantin,</font></div>
<div><font face="Calibri"></font> </div>
<div><font face="Calibri">I have deployed RabbitMQ cluster with three
nodes.</font></div>
<div><font face="Calibri">I wan to connect to any of the active nodes in case of
node failure.</font></div>
<div><font face="Calibri">I have tried TCP load balancer where I was
unsuccessful.</font></div>
<div><font face="Calibri">Could you please support me with some ideas how it can
be done from java client.</font></div>
<div><font face="Calibri">Thanks in advance.</font></div></div><div> </div>