The "maxmemory" setting is -Xmx, which is just the heap size. The JVM allocates 
more memory than that, such as for the permgen and the stack.

You might need to set the -XX:MaxPermSize parameter as well as the maxmemory. 
If you are on 64-bit, you can save some memory with -XX:+UseCompressedOops, or 
running in 32-bit mode.

However, depending on the amount of traffic you expect to do and the amount of 
data it will process, you might want to dedicate that server to repository 
management. 512m is about the right amount to adequately process any indexing 
and scanning tasks without spending a lot of time garbage collecting.

- Brett

On 29/05/2012, at 8:03 PM, Olivier Lamy wrote:

> Hi,
> 
> 
> 2012/5/29 Erwin Mueller <[email protected]>:
>> Hello,
>> how can I limit the memory usage for Archiva? I have only a 512MB RAM server
>> available, and Archiva is using 50% of the RAM.
>> 
>> I like to know how to set a hard limit on the RAM available for Archiva and
>> tips how I can minimize the RAM usage of Archiva.
>> 
>> I thought that the wrapper.conf can set a hard limit, but it's not working
>> (Archiva is using 514M virtual and 254M resources, but the limit is set to
>> 256M). I added/changed in wrapper.conf:
>> 
>> wrapper.java.additional.7=-Xms256m
>> wrapper.java.additional.8=-Xmx256m
>> ...
>> wrapper.java.initmemory=256
>> ...
>> wrapper.java.maxmemory=256
> 
> Yup that's normally the setup to change.
> Which version Archiva are you using ?
> How do you take your figures on memory usage ?
> 
>> 
>> Kind regards,
>> Erwin.
>> 
>> --
>> Erwin Mueller
>> http://www.mueller-public.de - My site;
>> http://www.global-scaling-institute.de/ - Global Scaling Institute GmbH;
>> http://www.anr-institute.com - Advanced Natural Research Institute GmbH;
> 
> 
> 
> -- 
> Olivier Lamy
> Talend: http://coders.talend.com
> http://twitter.com/olamy | http://linkedin.com/in/olamy

--
Brett Porter
[email protected]
http://brettporter.wordpress.com/
http://au.linkedin.com/in/brettporter
http://twitter.com/brettporter





Reply via email to