>> >I think this is a valid issue. I have only limited time right now, so can >> we work together on that? I'd do >> > instrumentation and you can run it? I think the environment with the easy >> repro is fine. >> >> Absolutely, I'll do whatever I can. >> > >Thanks! Let me first think about the best next step for instrumentation. I'll try to find it today. > >Quick question, though: on the environment where you can easily reproduce it: does it always carsh immediately (with the >first message)? Or does it need a couple of messages to crash. If the latter is the case, it would be great if you could >gather a debug log from the version I instrumented last time. That would enable me to see if the property interface >pointer is always invalid or gets smashed in course of processing...
It varies, but certainly more than one. I've attached valgrind logs of the imuxsockbug branch and for good measure, the current Debian package from the Adiscon APT repository. The test machine in this case was a new VM with the upstream Adiscon repositories added and the build and runtime dependencies installed from there (modulo the libgt packages: I alien'ed the RPM versions from the Adiscon RPM repository instead). The imuxsockbug branch was built on this machine with "--enable-memcheck --enable-valgrind --enable-debug --disable-testbench". I'll try to stick to building & testing on this machine from now on to keep things consistent. Interestingly though the stack trace has changed, which smacks of a stack smash. Again I'm able to crash using my logsmash tool but not with logger run in a tight loop. One difference I can think of is that logsmash keeps the socket open and continually calls syslog() in a tight loop, where-as my one liner Bash script naturally means that each logger instance is only ever sending a single message and dropping the socket. Let me know if there's anything else you need. -- Kristian Van Der Vliet
valgrind-logs.tar.gz
Description: GNU Zip compressed data
smime.p7s
Description: S/MIME cryptographic signature
_______________________________________________ rsyslog mailing list http://lists.adiscon.net/mailman/listinfo/rsyslog http://www.rsyslog.com/professional-services/ What's up with rsyslog? Follow https://twitter.com/rgerhards NOTE WELL: This is a PUBLIC mailing list, posts are ARCHIVED by a myriad of sites beyond our control. PLEASE UNSUBSCRIBE and DO NOT POST if you DON'T LIKE THAT.
