Andreas,
the log shows that some transactions cannot be deleted because some
referencing problem, I suspect it is somehow related to the ACK fix. I
did a small fixup on CVS and added some more debug in REF/UNREF to try
to figure out what is the problem.
So, first step is to update from cvs and re-run your tests - if the
problem persists, please send me the log again. I will also try to give
you a patch to revert the ACK changes to see if the problem disappears.
regards,
bogdan
Andreas Granig wrote:
Hi Bogdan,
could you please send me privately the shm mem log during run time
(SIGUSR1) and at shutdown? I will take a look and see if I can get
something out of it....
what it will be very, very , very helpful is the complete log (from
startup to shutdown) of openser with mem debug support in debug=9. It
will contain all logs about mem allocations and it will help me to
find what's the context of the leaks. This log can be huge I guess :-/
What you can find is a log of a freshly started openser with 100 calls
made. After that, I did a mem-dump with SIGUSR1, waited for ~45
minutes and shut it down.
What I'm curious about is what's about that 20 minutes waiting? I
couldn't see any mem-cleanup during the 45 minutes...
I've also attached the ser-config, if that helps you...
Thanks for your support,
Andy
_______________________________________________
Devel mailing list
[email protected]
http://openser.org/cgi-bin/mailman/listinfo/devel
