On 14.01.2012 23:58, Carl R. Friend wrote: > Greetings, > > Are there any other masochists out there who run monitoring systems > on anything other than Intel kit atop some form of GNU operating > system? If so, I am specifically curious about the ido2db executable > and how well it behaves. > > To set the stage, I'm running a Sun T1 105 with Solaris 10 Update 8, > gcc version 2.95.3, MySQL 5.0.51b, and Icinga 1.6.1.
gcc4 is preferred. gcc3 and sooner cause funny problems with realloc. how about sunfreeware tools? > Basic Icinga > works like a charm, but when I try lighting off ido2db, the latter > segfaults. The debug output indicates that it's likely happening in > the ido2db_db_query() routine (note that I had to hack the source to > remove the reference to the instance name; a structure-passing issue > perhaps): > > [1326578463.586885] [001.2] [pid=3049] [tid=1] ido2db_db_hello(25) conninfo_id > [1326578463.586973] [001.2] [pid=3049] [tid=1] ido2db_db_hello() get cached > object ids > [1326578463.587052] [001.2] [pid=3049] [tid=1] ido2db_get_cached_object_ids() > start > [1326578463.587179] [001.2] [pid=3049] [tid=1] ido2db_db_query() start > [1326578463.587593] [002.0] [pid=3049] [tid=1] SELECT object_id, > objecttype_id, name1, name2 FROM icinga_objects WHERE instance_id=0 instance_id=0 is most likely an error. can we see the full debug log for further reference? i would expect some libdbi function (get_uint or equal) to cause faulty memory alignments. which libdbi* versions are and installed and how? were there any warnings on configure? > > [1326578463.608127] [001.2] [pid=3049] [tid=1] ido2db_child_sighandler() start > [1326578463.608253] [001.2] [pid=3049] [tid=1] Child caught signal '11' > exiting > [1326578463.608463] [001.2] [pid=3049] [tid=1] ido2db_child_sighandler() end at this stage, get yourself a debugger. truss and gdb as well. read on how to run ido2db in foreground and analyse the error. https://wiki.icinga.org/display/Dev/Debug#Debug-IDOUtils > > > The output in /var/log/userlog is slightly less than useful: > > Jan 14 17:00:47 t1 ido2db[3049]: [ID 376905 user.info] IDO2DB 1.6.1 > (12-02-2011) Copyright(c) 2005-2008 Ethan Galstad, Copyright(c) 2009-2011 > Icinga Development Team (https://www.icinga.org) > Jan 14 17:00:47 t1 ido2db[3049]: [ID 255109 user.info] IDO2DB 1.6.1 > starting... (PID=3049) > Jan 14 17:01:03 t1 ido2db[3049]: [ID 973396 user.info] Client connected, data > available. > Jan 14 17:01:03 t1 ido2db[3049]: [ID 572053 user.info] Handling client > connection... > Jan 14 17:01:03 t1 ido2db[3049]: [ID 663190 user.info] Successfully connected > to mysql database > > whereupon it conveniently dies. If I'm running ido2db as a daemon, > it restarts at a different PID, but the outcome remains the same. > The icinga.log proper merely shows that the database connection has > gone away. that's rather normal as ido2db does not log to syslog when it segfaults. would have caused race conditions on participating threads. it's intentionally left in to the debug log to report such failures in deep for those actually interested in it. > > > The "icinga_instances" table does get populated, as does the > "icinga_conninfo" one, but nothing else. populated how? > > > I have not tried this on an Intel platform, but suspect that it all > "just works" there as other folks seem to be having luck with it. I > recall putting up a big fight with ndoutils-1.4b9, eventually getting > it to work, but the ido2db code looks like the problems I found in > ndoutils had been fixed. ndo2db uses plain mysql-client api, ido2db abstracts that layer to libdbi. i would expect an error in between. maybe due to an old compiler used. > > > Ideas? C code should "just work" no matter what platform it's > compiled on, so there may be hidden issues in the mix as well. solaris is and was always special and a pita to debug for errors. i've debugged the core myself while getting explained by a colleague how to work with solaris, leaving 2 weeks of my life for some gcc3 issue. not one of my favorite systems now - unless i would have endless time in my life. kind regards, michael > > > Cheers! > > +------------------------------------------------+---------------------+ > | Carl Richard Friend (UNIX Sysadmin) | West Boylston | > | Minicomputer Collector / Enthusiast | Massachusetts, USA | > | mailto:crfri...@rcn.com +---------------------+ > | http://users.rcn.com/crfriend/museum | ICBM: 42:22N 71:47W | > +------------------------------------------------+---------------------+ > > ------------------------------------------------------------------------------ > RSA(R) Conference 2012 > Mar 27 - Feb 2 > Save $400 by Jan. 27 > Register now! > http://p.sf.net/sfu/rsa-sfdev2dev2 > _______________________________________________ > icinga-users mailing list > icinga-users@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/icinga-users -- DI (FH) Michael Friedrich Vienna University Computer Center Universitaetsstrasse 7 A-1010 Vienna, Austria email: michael.friedr...@univie.ac.at phone: +43 1 4277 14359 mobile: +43 664 60277 14359 fax: +43 1 4277 14338 web: http://www.univie.ac.at/zid http://www.aco.net Lead Icinga Core Developer http://www.icinga.org ------------------------------------------------------------------------------ RSA(R) Conference 2012 Mar 27 - Feb 2 Save $400 by Jan. 27 Register now! http://p.sf.net/sfu/rsa-sfdev2dev2 _______________________________________________ icinga-users mailing list icinga-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/icinga-users
