I'm running syslog-ng 1.4.7 on a Solaris 7 machine which acts as a middleman for my production servers and main log repository server. It crashes without warning, without logging any problems beforehand, and often requires several attempts at restarts before staying up. Time between crashes seems random; it will go for a few days, die, require ten restarts in five minutes to keep it up, then run for another week before having any problems. 'gdb' says it's dying from a SIGKILL but I don't have anything running on the machine that would send it any such flags. Any ideas what is going on here? I found two messages in the archives from Oct. 2001 asking similar questions, but no resolutions: Nicholas Berry nberry@ragingwire.com Wed, 10 Oct 2001 09:10:13 -0700 a.. I too am having the same issue with crashing under Solaris 8 with syslog-ng 1.5.10, but my crashes happen several hours between each other. _________________________________________________________________ Nicholas Berry RagingWire Telecommunications, Inc. Systems Engineer 2710 Gateway Oaks-South, Suite 300 nberry@ragingwire.com Sacramento, CA 95833 http://www.ragingwire.com Tel: 916.286.4048 Fax: 916.921.4048 -----Original Message----- From: Andreas Schulze [mailto:Andreas.Schulze@mediaWays.NET] Sent: Wednesday, October 10, 2001 6:01 AM To: syslog-ng list Subject: [syslog-ng]Mystic crashes of 1.5.10 Hello Bazsi, thanks a lot for the stats() feature in the new version. But unfortunately 1.5.10 crashes a few minutes after starting. We are running your software on SUN Solaris 8 boxes. I read about a similiar problem with Linux boxes some weeks ago in this list. For your convenience the backtrace is attached. The calling deps between #17 and #1 can differ in each core. But it crashed _always_ in __sigprocmask () from /usr/lib/libthread.so.1 And it crashed _always_ by receiving SIGKILL. Seems a little bit mystic for me, because I don't know why and who's sending the SIGKILL. Coredumps by SIGSEGV are more familiar to me :) Thx. for a fast fix --Andreas