hmm, didn't Fred say he gets his problem using the commandline request
and insert clients?  We need to get straight on whether Fred's client
CPU suck has anything to do with runaway threads, and on how exactly these
possibly separate problems are reproduced..

apparently all we know right now is Kirk seems to get runaway threads
reproducibly from failed fproxy requests.  Does he get them from
failed commandline requests?  Does Fred get his problem from fproxy,
freenet_request, or both?  Is it tied to failed requests?

I'm not sure whether fproxy could be run in a separate jvm, but the
commandline clients are, aren't they?


On Fri, Feb 23, 2001 at 11:39:13AM -0500, Chris Anderson wrote:
> 
> Tavin & Scott are probably right, Fred & Kirk may be seeing lost threads
> in fproxy.  This would be useful knowledge, so if it's possible run the
> proxy in a separate jvm and you should be able to tell which process owns
> the runaway threads.  On linux, a command like this:
> 
> $ ps -eo pid,ppid,pgrp,args
> 
> will show the parent process id and the process group id.  I think win2k
> task manager is able to devine this info as well.
> 
> In Kirk's top output, it looked like there were 2 groups of 9 threads, 2
> groups of 4 threads and a couple of stragglers.  I suspect the 2 groups of
> 4 are fproxy threads... 1-listener, 1-timer, 1-connection handler, 
> 1-client.
> 
> 
> _______________________________________________
> Devl mailing list
> Devl at freenetproject.org
> http://www.uprizer.com/mailman/listinfo/devl

-- 

/* tavin cole
 * composer of e-mail messages
 */

_______________________________________________
Devl mailing list
Devl at freenetproject.org
http://www.uprizer.com/mailman/listinfo/devl

Reply via email to