Mauricio Nuñez wrote:
El jue, 28-07-2005 a las 14:25 +0200, Remy Maucherat escribió:
OK. But may be more easy, It's not very "out of the box" .
Since the out of the box version has a bug, the "out of the box"
experience is going to suck, there's nothing I can do about it.
I noticed that after trying random stuff. While it's good to send
details, it's less useful to send them after the bug has been fixed.
OK. The test case is for the list archive, not only for you. The smart
people search the archives before send a bug.
Ok.
I'm testing your change, all is working again.
65 ajp workers on machine 1, 95 ajp workers on machine 2.
The heap usage grow smoothly.
My sample is about 20 min.
What kind of data may be relevant for you ?
That it's working is pretty good already, but you should mention your
OS. I built the AJP APR implementation using lots of cut & paste of the
existing AJP code, and for the most part, just to "have it", as I don't
expect major benefits (except in one special case where the amount of
request processors on the front end is huge compared to the amount of
request processors that can realistically be used on the back end -
often it's the other way around, so there's no problem).
Rémy
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
