I have a radiant instance whose memory footprint grows to 200+ megs
of RAM. At which time I have to restart the service. Does anybody
have any information about how to size memory requirements for
Radiant? (For example, a lot of J2EE servers require 512 Megs to do
anything useful, but can live quite well within a Gig.)
If it makes a difference, I’m running Radiant 0.5.2 on windows with
Rails 1.1.6, backed by SQL Server, running in Production mode… It’s
behind an IIS server proxy. Part of the memory headache seems to be
Windows related, as the memory tops out and “interesting” behaviors
begin to occur.
Sincerely,
Paul H.
On 2/5/07, Paul H. [email protected] wrote:
begin to occur.
Hello Paul,
How are you running your ruby process? Using Webrick and a ISAPI proxy?
Which version of ruby are you suing? 1.8.4 or 1.8.5?
In .5 exist a method to control the memory use of your process:
http://www.rubyinside.com/control-your-rails-process-sizes-with-ruby-185-246.html
Also, I’ll suggest you look into Mongrel and mongrel_service
http://mongrel.rubyforge.org
A complex, big radiant website (400 pages) will not use more than
100MB of RAM per process (in case you’re running a cluster).
Also, current Ruby implementation show memory leaks in long-running
process that use threads, mutexes and synchronization mechanism.
I’ll suggest install and use FastThread from MenTaLGuY (gem install
fastthread) and require it early in your config/environment.rb script.
Hope these tips help you,
Regards,
–
Luis L.
Multimedia systems
Leaders are made, they are not born. They are made by hard effort,
which is the price which all of us must pay to achieve any goal that
is worthwhile.
Vince Lombardi
Thanks, Luis, I’ll check out the links you sent. Here’s the rest of
the information:
I’m running the process using mongrel. Radiant is configured as a
windows service. The proxy is a combination of the rewriter from
Helicon - ISAPI_Rewrite and IIS. It just proxies the request to the
windows service, running on a local only network connection. (So you
have to go directly to radiant). It’s not a large site with pages is
the dozens - not hundreds.
On 2/5/07, Paul H. [email protected] wrote:
Thanks, Luis, I’ll check out the links you sent. Here’s the rest of
the information:
I’m running the process using mongrel. Radiant is configured as a
windows service.
I’ll suggest you make sure the params used by the service (Check the
Path to Executable shown on Services in Administrative Tools).
Mostly because running Radiant in development mode don’t activate the
caching engine and make rails reload all the clases on each request,
something that will rise memory leaking issues.
The proxy is a combination of the rewriter from
Helicon - ISAPI_Rewrite and IIS. It just proxies the request to the
windows service, running on a local only network connection. (So you
have to go directly to radiant). It’s not a large site with pages is
the dozens - not hundreds.
Ok, good to know you’re on mongrel 
If none of the suggestions I droped previously work, please post this
case at mongrel-users mailing list: [email protected]
http://rubyforge.org/mailman/listinfo/mongrel-users
Rails 1.1.6, backed by SQL Server, running in Production mode… It’s
Which version of ruby are you suing? 1.8.4 or 1.8.5?
A complex, big radiant website (400 pages) will not use more than
Regards,
Radiant mailing list
Post: [email protected]
Search: http://radiantcms.org/mailing-list/search/
Site: http://lists.radiantcms.org/mailman/listinfo/radiant
Radiant mailing list
Post: [email protected]
Search: http://radiantcms.org/mailing-list/search/
Site: http://lists.radiantcms.org/mailman/listinfo/radiant
–
Luis L.
Multimedia systems
Leaders are made, they are not born. They are made by hard effort,
which is the price which all of us must pay to achieve any goal that
is worthwhile.
Vince Lombardi