Very well explained Peter. Thanks a lot. I take a note about the translation.
;)
Peter Crowther wrote:
>
>> From: HLL [mailto:vicretru@gmail.com]
>> I have a doubt. How can I determine the optimal number of
>> Tomcat servers in
>> the same machine to achive the best throughput depending on
>> the number of
>> cores of the machine? That is, which is the optimal relation
>> between number of tomcats and number of cores.
>
> Measure your application, under your expected load, on your hardware.
> There is no other way.
>
> If your application is well optimised and is not waiting on locks, the
> optimal number of Tomcats may be 1. This optimises the cache behaviour of
> the cores, as you don't have multiple copies of the same code and data
> loaded at different locations and competing for cache lines. Lock
> contention, garbage collector behaviour or many other variables could
> change that, and make it more optimal to segment your application. If
> your machine is large enough to have a memory architecture where some
> processors have faster access to some memory, the optimal number of
> Tomcats may be one per memory segment - but this will need a *lot* of
> analysis and tuning.
>
> [For future posts to forums: A better English translation may be
> "question" rather than "doubt"]
>
> - Peter
>
> ---------------------------------------------------------------------
> To start a new topic, e-mail: users@tomcat.apache.org
> To unsubscribe, e-mail: users-unsubscribe@tomcat.apache.org
> For additional commands, e-mail: users-help@tomcat.apache.org
>
>
>
--
View this message in context: http://www.nabble.com/Optimal-number-of-Tomcat-instances-versus-Number-of-cores-tp17228866p17229456.html
Sent from the Tomcat - User mailing list archive at Nabble.com.
---------------------------------------------------------------------
To start a new topic, e-mail: users@tomcat.apache.org
To unsubscribe, e-mail: users-unsubscribe@tomcat.apache.org
For additional commands, e-mail: users-help@tomcat.apache.org
|