On Wed, May 15, 2002 at 01:15:41PM -0400, John Morrissey wrote:
On Wed, May 15, 2002 at 10:20:45AM +0200, Balazs Scheidler wrote: % sorry to repeat myself, but this patch adds even more sanity checking:
I patched libol on a local machine and started syslog-ng. It ran for much longer than normal, but then I realized that it could reach the syslog hub. When I removed its access to the syslog hub, syslog-ng died (segfaulted) within a couple minutes with the following internal messages logged:
May 15 12:44:09 serviceman01 syslog-ng[9566]: syslog-ng version 1.4.14 starting May 15 12:44:09 serviceman01 syslog-ng: syslog-ng startup succeeded May 15 12:44:50 serviceman01 syslog-ng[9566]: Connection broken, reopening in 60 seconds May 15 12:49:22 serviceman01 syslog-ng[9566]: Connection broken, reopening in 60 seconds
Backtrace was the same. There were no assertions logged; would they be syslogged? It looks like syslog-ng gets wedged somehow before it dies and might not be able to log anything.
the asserts wouldn't be syslogged, but the terminating signal would become SIGABRT instead of SIGSEGV. 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. How did you remove its ability to send messages to the hub? using a packet filter rule, directing it to an unused port, or? -- Bazsi PGP info: KeyID 9AF8D0A9 Fingerprint CD27 CFB0 802C 0944 9CFD 804E C82C 8EB1