2013/6/25 Ronald Moesbergen <intercom...@gmail.com> > 2013/6/24 Paul J Stevens <p...@nfg.nl> > >> >> You /could/ try >> >> >> diff --git a/src/clientsession.c b/src/clientsession.c >> index b59c8ee..f8b670b 100644 >> --- a/src/clientsession.c >> +++ b/src/clientsession.c >> @@ -138,6 +138,8 @@ void client_session_bailout(ClientSession_T **session) >> List_T messagelst = NULL; >> >> if (! c) return; >> + ci_cork(c->ci); >> + >> TRACE(TRACE_DEBUG,"[%p]", c); >> // brute force: >> if (server_conf->no_daemonize == 1) _exit(0); >> > > I tried this patch, but it doesn't help :( I still managed to crash the > lmtp daemon: > > Program received signal SIGSEGV, Segmentation fault. > 0x00007f2cbeca29a7 in client_session_set_timeout (session=0x939ea0, > timeout=60) at clientsession.c:243 > 243 session->ci->timeout->tv_sec = timeout; > (gdb) bt > #0 0x00007f2cbeca29a7 in client_session_set_timeout (session=0x939ea0, > timeout=60) at clientsession.c:243 > #1 0x00007f2cbeca2b6a in socket_write_cb (fd=18, what=4, arg=0x939ea0) at > clientsession.c:273 > #2 0x00007f2cbcfd294c in event_base_loop () from > /usr/lib/libevent-2.0.so.5 > #3 0x00007f2cbeca0a66 in server_run (conf=0x7fff2dfd7210) at server.c:823 > #4 0x00007f2cbeca1023 in server_mainloop (config=0x7fff2dfd7210, > service=0x4048a5 "LMTP", servicename=0x4048ed "dbmail-lmtpd") at > server.c:957 > #5 0x0000000000402bcb in main (argc=1, argv=0x7fff2dfdab78) at lmtpd.c:48 > > > Lowering concurrency on the client side is possible, I'll see if that > works around the issue. But it will still be just that: a workaround. I'm > not really comfortable using this in production knowing this bug still > exists. > > I've checked with lower concurrency (2 thread per sending server) and that seems to work, at least in my test setup. I've also found another workaround: run dbmail-lmtpd under 'xinetd', using dbmail-lmptd in stdin/stdout mode (-n). This way there's always one client talking to one dbmail-lmtpd process. An added bonus when using this method is that it's much faster and can handle much more traffic. Also, there's no long-lived dbmail-lmtpd process that causes downtime when it crashes...
Thanks, Ronald.
_______________________________________________ Dbmail-dev mailing list Dbmail-dev@dbmail.org http://mailman.fastxs.nl/cgi-bin/mailman/listinfo/dbmail-dev
