On 06/26/2012 07:45 AM, roy.ja...@xemaps.com wrote:
Hello James devs!
So our custom implementation of James is truckin' along and we started running
load tests. Despite hundreds of connections I can't seem to get James' IMAP
service to utilize more than 60% (usually hovers around 50%) cpu. This could
be due to our custom implementation but I was wondering if there were any
parameters I could set that could help me tune the service?
Glad to read this.
Btw, which tool do you use to stress on IMAP?
I do see ioWorkerCount and maxExecutorCount but these don't seem to change CPU
usage for me. Any suggestions would be great!
It's possible that the limit resides on the IO (disk and/or network).
What does iostat tell? (iostat -x 1)
Unfortunately I have yet to get around to figuring out what was wrong with the
COMPRESS option for IMAP. Partially due to lack of time (we can deal with no
compression for now) and partially due to my newbiness. I tried to grok the
code for COMPRESS but was left kind of confused. I tried to follow the chain
of command but for some reason could not find where the compression was
actually handled. Could someone point out the class that actually does the
compress and decompress?
I still have your build which breaks the ipad.
I made test and it was indeed giving issue on my ipad, but I didn't
further investigate.
Regards,
Roy
---------------------------------------------------------------------
To unsubscribe, e-mail: server-dev-unsubscr...@james.apache.org
For additional commands, e-mail: server-dev-h...@james.apache.org
--
eric | http://about.echarles.net | @echarles
---------------------------------------------------------------------
To unsubscribe, e-mail: server-dev-unsubscr...@james.apache.org
For additional commands, e-mail: server-dev-h...@james.apache.org
