> Looking at the log, it looks like MSSQL is continously failing as long as the log > runs. Rsyslog retries, then tries to begin a new transaction and then MSSQL > returns an error state (at least this is what it looks ;)).
That is interesting. I cannot see any tcp SYN packet, once the TCP connection is terminated. > Can you see anything in some MSSQL tools to see if there is a new connect? I'm not the MSSQL guy, but I will ask the colleague once again. Frankly, I doubt that there is any new connect, because I cannot see it with tcpdump. > It looks like we have one of three options > > a) bug in rsyslog's omlibdbi > b) problem in libdbi's MSSQL driver and reconnect > c) problem with MSSQL itself > > I hope it's a) and will have some more look at the code. I'll probably create a > special version for your use. If so, can you build from source and execute that > version? I am asking because without that, creating instrumented code makes > little sense ;) Yes, I can build from source, actually I compiled libdbi, libdbi-drivers, libee, librelp, libestr and rsyslog from source. Freetds came from the package manager: freetds-0.91-2.el6.x86_64 freetds-devel-0.91-2.el6.x86_64 > Do you know valgrind? Could you run your lab under valgrind control? > That usally is immensely helpful. Now I have an additional testing machine with valgrind compiled and running. What do you want me to do exactly? Martin
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.
