On Wed, 2 Aug 2000, Scott McDermott wrote:
Balazs Scheidler on Sat 29/07 19:45 +0200:
I'm running syslog-ng on a FreeBSD 3.2-RELEASE machine. It does work but tends to start using alot of CPU time and after that stops logging. Maybe it's my dodgy config...
I'm not using syslog-ng on FreeBSD. Could you provide me some more details? ktrace dump would be helpful.
Incidentally, the 1.4.5 release did fix some of the leak/CPU problems, but only made it a longer period of time before the process eventually uses all system resources; there's still a leak somewhere. I'll have more details within the next couple of weeks.
I've had lots of problems with the stability of syslog-ng on two debian installs, but look at the syslog-ng listing from top on a redhat 6.2 box: PID USER PRI NI SIZE RSS SHARE STAT LIB %CPU %MEM TIME COMMAND 355 root 11 0 20952 20M 432 R 0 56.7 2.6 12616m syslog-ng It's been up and running for `expr 12616 / 60 / 24` == about 8 days with no problems, and it's logging both to the regular logfiles in /var/log plus archiving each host like this: # separate logs destination std { file("/var/log/HOSTS/$HOST/$FACILITY/$YEAR$MONTH/$FACILITY$YEAR$MONTH$DAY" owner(root) group(root) perm(0600) dir_perm(0700) create_dirs(yes)); }; # log it log { source(net); source(local); destination(std); }; This is for around 50 hosts, so it's a not excactly a light load. It consumes most of the CPU since it's usually the only running process. This is syslog-ng 1.4.4, I forget which libol version (deleted the tarball). I wonder what's so magic about redhat? Any thoughts? -- Nate