> -----Original Message----- > From: [email protected] [mailto:rsyslog- > [email protected]] On Behalf Of John Feuerstein > Sent: Tuesday, November 09, 2010 12:32 AM > To: rsyslog-users > Subject: Re: [rsyslog] One thread eating up infinite CPU time > (spinlock?) > > Hello, > > On 11/08/2010 08:47 PM, [email protected] wrote: > > can you get the lsof output so that we can see what the file > descriptor > > that it's having problems with is supposed to be? > > I've attached strace for about 10 seconds using: > > $ strace -ff -v -p 32267 -o rsyslogd > > ... and while strace was running, I've captured lsof output using: > > lsof -n -p 32267 > rsyslogd.lsof > > You can find all files and a tarball of them here: > http://biz.baze.de/debug/rsyslog/ > > Hope this helps. FYI, I've also run this for about 10 seconds on the > problematic thread: > > $ strace -p 32270 -c > Process 32270 attached - interrupt to quit > Process 32270 detached > % time seconds usecs/call calls errors syscall > ------ ----------- ----------- --------- --------- ---------------- > 60.95 0.003239 0 30405 select > 36.24 0.001926 0 1033772 1033709 recvfrom > 2.80 0.000149 0 30405 30405 accept > 0.00 0.000000 0 64 sendto > 0.00 0.000000 0 124 futex > ------ ----------- ----------- --------- --------- ---------------- > 100.00 0.005314 1094770 1064114 total
which inputs do you run? Only imrelp? The bottom line is that we need to actually see what rsyslog is doing. If you version is not too old, we could use debug on demand (in the link, Florian says he tested it with 6.x, but left out the more interesting fact of which is the minimum version ;)): http://www.rsyslog.com/how-to-use-debug-on-demand/ Rainer _______________________________________________ rsyslog mailing list http://lists.adiscon.net/mailman/listinfo/rsyslog http://www.rsyslog.com
