[rabbitmq-discuss] rabbitmq java client socket io performance (3 sendto syscalls per basicPublish)

Jason Toffaletti jason at cornsyrup.org
Thu May 24 09:39:38 BST 2012


I was looking at the code in hg that includes this patch. Now it makes sense why I couldnt reconcile the behavior with the code. I'll look forward to the next release and test again then.

Thanks

On May 24, 2012, at 1:22 AM, Matthias Radestock <matthias at rabbitmq.com> wrote:

> On 24/05/12 08:46, Jason Toffaletti wrote:
>> I was stracing a java process that uses the 2.8.2 java amqp client and I
>> noticed that every basicPublish is resulting in 3 sendto syscalls:
>> [...]
> 
> https://github.com/rabbitmq/rabbitmq-java-client/commit/41c67bf0502d4a13d08c4f42dc47721c5048779f changes that and will be in the next release.
> 
> The change actually makes very little difference to performance unless its an SSL connection.
> 
> Matthias.


More information about the rabbitmq-discuss mailing list