> W dniu 04.06.2012 12:02, Rainer Gerhards pisze:
> > Marcin,
> >
> > can you pls try this patch:
> >
> >
> http://git.adiscon.com/?p=rsyslog.git;a=commitdiff;h=55659b96a38d926445c9
> ae4123df4b1d29e18c32
> >
> > It worked for me, but that doesn't say anything because it also works
> > without
> that patch (I don't see the race...).
>
> I tried patch but i can't any see diffrence. I can see only "start" in
> log. I'm attaching debug log (maybe it helps).
Strange...
Can you add the patch below and send me a new debug log. Default
instrumentation does not convey enough info. Also, I have now extended the
timeout quite a bit...
Rainer
diff --git a/tools/syslogd.c b/tools/syslogd.c
index 0988d46..dddde48 100644
--- a/tools/syslogd.c
+++ b/tools/syslogd.c
@@ -1065,6 +1065,7 @@ die(int sig)
/* and THEN send the termination log message (see long comment above) */
if(sig && bLogStatusMsgs) {
+DBGPRINTF("logging exit message...\n");
(void) snprintf(buf, sizeof(buf) / sizeof(char),
" [origin software=\"rsyslogd\" " "swVersion=\"" VERSION \
"\" x-pid=\"%d\" x-info=\"http://www.rsyslog.com\"]"; " exiting
@@ -1076,7 +1077,7 @@ die(int sig)
* otherwise we have seen cases where the message did not make it to log
* files, even on idle systems.
*/
- srSleep(0, 50);
+ srSleep(2, 50);
/* drain queue (if configured so) and stop main queue worker thread pool
DBGPRINTF("Terminating main queue...\n");
_______________________________________________
rsyslog mailing list
http://lists.adiscon.net/mailman/listinfo/rsyslog
http://www.rsyslog.com/professional-services/
What's up with rsyslog? Follow https://twitter.com/rgerhards
