On Thu, May 16, 2002 at 11:57:28AM +0200, Balazs Scheidler wrote: % the asserts wouldn't be syslogged, but the terminating signal would become % SIGABRT instead of SIGSEGV. Still looks like SIGSEGV: [root@villa:pts/1 /staff/jwm> gdb `which syslog-ng` core [snip] Core was generated by `syslog-ng'. Program terminated with signal 11, Segmentation fault. % I'm trying to reproduce the problem here. I changed the reopen time from 60 % to 3 seconds, and sending log messages using logger with a shell script. If you want direct access to a machine, I might be able to get it for you. % How did you remove its ability to send messages to the hub? using a packet % filter rule, directing it to an unused port, or? A filter rule on the hub that would cause the client to receive ECONNREFUSED. I don't know if it helps, but I think it's interesting that if I start syslog-ng with the hub "unavailable," it seems to run for a a much longer time (properly?), but if I start syslog-ng and *then* remove access to the hub, syslog-ng dies within a few minutes. john -- John Morrissey _o /\ ---- __o jwm@horde.net _-< \_ / \ ---- < \, www.horde.net/ __(_)/_(_)________/ \_______(_) /_(_)__