July 2006 - Xymon - Xymon mailman web

bb-xsnmp problem
by damiano.vozza＠apss.tn.it 14 Jul '06

14 Jul '06

I use Hobbit 4.1.2.p1 on FC4 and I tried to test bb-xsnmp.pl v. 1.79 with a Cisco router. I ran the test on command line with /usr/local/hobbit/server/bin/bbcmd bb-xsnmp.pl and all worked fine. I tried to launch it from hobbitlaunch [xsnmp] ENVFILE /usr/local/hobbit/server/etc/hobbitserver.cfg NEEDS hobbitd CMD /usr/local/hobbit/server/ext/bb-xsnmp.pl LOGFILE $BBSERVERLOGS/bb-xsnmp.log INTERVAL 5m but hobbitlaunch process used a lot of VM and the machine slows down; I had to stop Hobbit . No bb-xsnmp.log was produced. Some help? Damiano Vozza Azienda Provinciale per i Servizi Sanitari Servizio Sistemi Informativi Ospedale S. Chiara Via Crosina Sartori 6 Trento - Italy tel. (+39) 0461364856 fax. (+39) 0461903593 e-mail damiano.vozza at apss.tn.it <mailto:damiano.vozza at apss.tn.it> web site www.apss.tn.it <http://www.apss.tn.it/>

1 0

[hobbit] Error in hobbitdboard request crashes hobbitd
by tschenk＠pironet-ndh.com 14 Jul '06

14 Jul '06

> > > > I am testing the 4.2-RC on Linux 2.6.15. Trying all the new features > > (thanks, henrik!) I sent a hobbitdboard request with an error in it: > > > > bb $BBDISP "hobbitdboard status=yellow" > > (should have been color=yellow) > > > > This causes the hobbitd process to crash with a segmentation fault. > > Thanks, this is indeed a bug. A patch is available on the betapatches > page at http://www.hswn.dk/hobbitsw/betapatches/ > Applied the patch, works good now. Thanks, Thomas

1 0

Error in hobbitdboard request crashes hobbitd
by tschenk＠pironet-ndh.com 14 Jul '06

14 Jul '06

Hello, I am testing the 4.2-RC on Linux 2.6.15. Trying all the new features (thanks, henrik!) I sent a hobbitdboard request with an error in it: bb $BBDISP "hobbitdboard status=yellow" (should have been color=yellow) This causes the hobbitd process to crash with a segmentation fault. This is the gdb output of the core file: #0 0xb7ea983b in raise () from /lib/tls/libc.so.6 #1 0xb7eaafa2 in abort () from /lib/tls/libc.so.6 #2 0x0805baf1 in sigsegv_handler (signum=0) at sig.c:57 #3 <signal handler called> #4 0xb7ef9363 in strlen () from /lib/tls/libc.so.6 #5 0x0804e496 in setup_filter (buf=0x8495cd0 "hobbitdboard", defaultfields=0x805fde0 "hostname,testname,color,flags,lastchange,logtime,validtime,acktime,disa bletime,sender,cookie,line1", spage=0xbffdb404, shost=0xbffdb408, snet=0xbffdb40c, stest=0xbffdb410, scolor=0xbffdb414, acklevel=0xbffdb418, fields=0xbffdb41c, chspage=0xbffdb420, chshost=0xbffdb424, chsnet=0xbffdb428, chstest=0xbffdb42c) at hobbitd.c:2066 #6 0x0805117c in do_message (msg=0x82d86a0, origin=0x805e363 "") at hobbitd.c:2753 #7 0x08054478 in main (argc=7, argv=0xbffdf004) at hobbitd.c:4340 Sending this request from any bb or hobbit client seems to crash hobbitd. Regards, Thomas

2 1

A funny thing happened on the way to the RC...............
by brodie＠mcw.edu 14 Jul '06

14 Jul '06

So, I updated my server to the RC version today. Saved all of my configuration files, and (based on info from this list!), decided to add fping to my system this time. Fping installed, no problem. Fired up BB. Then, everyone's pager went NUTS. 100 pages, times 4 pagers. WHOOOOPS. The entire world had uh..... "lost connectivity." What had really happened of course is, that the hobbit userid couldn't run fping, since I had forgotten to suid the executable. And yes, I did eventually find it in a log file, but... But, this had the very nasty side effect of turning each and every connectivity icon RED. And then the alerts kicked in..... And four Duracell batteries died. I don't know how easy it is to implement, but I have a suggestion worth looking into? If fping "doesn't work" (as opposed to, "cannot ping a host"), the result status should be CLEAR, not RED. "test not functional" or "test disabled" or something. Just a thought!!!! :-) --------------------------------------------------------- Kent C. Brodie - brodie at phys.mcw.edu Department of Physiology Medical College of Wisconsin (414) 456-8590

2 1

[hobbit] RC client release bug?
by David.Gore＠verizonbusiness.com 14 Jul '06

14 Jul '06

Hello, We have seen this with recent snapshots and the latest release candidate client. logfetch hangs which causes the client to hang and go purple for all the tests. It can be resolved by killing logfetch and deleting all the entries in ~/client/tmp. We could try to be more surgical on the deleting of files. This has happened on two very independent hosts running Solaris 8, one being a SunFire 880 and another being an E4500/E5500. Suggestions? It can run for many days before hanging. -- David

2 6

Multiple User Admin mode
by steve.aiello＠ge.com 13 Jul '06

13 Jul '06

Today I am using BigBrother/BBgen 3.6. One of the custom features I have hacked into our system is the ability to have multiple admin users that can put servers into maintenance mode. In our BB configuration we made extensive use of the the HostGroup function. Basically creating an aliasname for a list of devices. This came in handy when writing alert rules, I could do one alert rule for a series of devices. Now for maintenance mode, I hacked the maint.pl script found on deadcat. That script was then put into a directory that apache password protected. When you logged in, the maint.pl script compared your login name to any HostGroup that was defined, and if the login name matched a HostGroup the maintenace device listing only listed the devices that were defined by the the HostGroup. Now I did see the archived mailing list article, Re: [hobbit] New Features #2 (the 'publish-cookie' option) (http://www.hswn.dk/hobbiton/2005/05/msg00369.html) That solution really only limits users from viewing different pages. The maint option under the device info report is located under the /hobbit-cgi/ and I do not see a way to limit users from disabling devices they shouldn't. If a user has an admin password, they seem to have admin rights to all of hobbit. Under Hobbit I do not see anyway to define HostGroups, so my old method will not work. Using HostGroupings seemed ideal in that it linked alerts being sent to users and the user's ability to disable devices/tests. ~Steve

2 1

Migration approach
by jeffnewman75＠gmail.com 13 Jul '06

13 Jul '06

All, I was thinking now that the new version of hobbit has moved on to RC (yay!) I would migrate my existing production hobbit server to the new version. What sort of migration approach is recommended? I ask this because I was thinking that it might be better to start fresh with the RRD's because the new version adds some that didn't exist before, and might do things differently with some of the existing ones, etc etc etc.. However, I don't want to lose the historical data, MEANING, I don't have the requirement to re-add the existing data, but would like to be able to refer to it. For example: The new hobbit is running for a week, and someone has XYZ problem, wants me to look at the graphs. I see there is some huge spike at xyz time, but I don't know how long it has been going on, or if this is a growing trend, or whatever, so I want to look at the "old hobbit" graphs for that host which has much more data. Is there any way to "trick" hobbit into looking at the old RRD files for a given host, or is there some approach I am missing? Am I just being dumb? <grin> Comments apprieciated. Thanks, jeff

2 1

Question about CLASS= & classname
by steve.aiello＠ge.com 13 Jul '06

13 Jul '06

Per the hobbit-clients.cfg documentation I read the following: CLASS=classname Rule match by the client class-name. You specify the class-name for a host when starting the client through the "--class=NAME" option to the runclient.sh script. If no class is specified, the host by default goes into a class named by the operating system. My questions: The default classname setting of the operating system, which value is it ? When I look at the client data of a host I see the following: [uname] Linux serverName 2.6.16-gentoo-r7 i686 [osversion] Gentoo Base System version 1.6.14 Is the default value of classname set to [uname] or [osversion]. Also is it the full name ? If so do I define my CLASS="Linux serverName 2.6.16-gentoo-r7 i686" ? Can regular expresion be used, CLASS=%Linux.* ? Can a host be defined to more than one class, i.e. 1.2.3.4 serverName # ssh CLASS:Linux CLASS:webServer If a host can have more than one class definition, what happens when you get CLASS configuration parameter collision ? Also, it may bee helpful to have the Info report show which class the device falls under. Thank you, Steve

1 0

[hobbit] How to totally disable alerts for a particular host?
by brodie＠mcw.edu 13 Jul '06

13 Jul '06

In truth, I think my issue is I'm not completely understanding the "perl compatable" regular expressions. Can someone perhaps include some examples? I am finding that some of my attempts below are actually causing ALL alerts to simply cease-- with nothing in the log files to indicate I have an issue!!! -----Original Message----- From: Brodie, Kent [mailto:brodie at mcw.edu] Sent: Wednesday, July 12, 2006 10:20 PM To: hobbit at hswn.dk Subject: [hobbit] How to totally disable alerts for a particular host? I'm not sure I understand the docs here... is something like this correct? # These rules eliminate printers from warnings/alerts ##HOST=%^pr*.hmgc.mcw.edu ## IGNORE HOST=lp4.phys.mcw.edu IGNORE ##HOST=%^hp* ## IGNORE # Misc hosts we do not want alerts on ##HOST=mac1.hmgc.mcw.edu ## IGNORE ##HOST=canadeo-disk.brc.mcw.edu ## IGNORE ##HOST=sharpe.phys.mcw.edu ## IGNORE HOST=bc1s7.phys.mcw.edu IGNORE # # Now on to our regular monitoring rules for things we DO want alerts on # HOST=uram.hmgc.mcw.edu MAIL brodie at phys.mcw.edu REPEAT=30 RECOVERED FORMAT=TEXT MAIL kent.brodie at pager.mcw.edu REPEAT=30 RECOVERED FORMAT=SMS ##SERVICE=disk ## MAIL brodie at phys.mcw.edu REPEAT=30 RECOVERED FORMAT=TEXT HOST=loki.phys.mcw.edu SERVICE=disk MAIL brodie at phys.mcw.edu REPEAT=60 RECOVERED FORMAT=TEXT SERVICE=conn MAIL sysadmins REPEAT=30 RECOVERED FORMAT=TEXT MAIL pagers REPEAT=30 RECOVERED FORMAT=SMS SERVICE=oracle MAIL szacher at hmgc.mcw.edu REPEAT=30 RECOVERED FORMAT=TEXT MAIL sysadmins REPEAT=30 RECOVERED FORMAT=TEXT MAIL kent.brodie at pager.mcw.edu REPEAT=30 RECOVERED FORMAT=SMS [hobbit at coffman etc]$ To unsubscribe from the hobbit list, send an e-mail to hobbit-unsubscribe at hswn.dk

2 1

How to totally disable alerts for a particular host?
by brodie＠mcw.edu 13 Jul '06

13 Jul '06

I'm not sure I understand the docs here... is something like this correct? # These rules eliminate printers from warnings/alerts ##HOST=%^pr*.hmgc.mcw.edu ## IGNORE HOST=lp4.phys.mcw.edu IGNORE ##HOST=%^hp* ## IGNORE # Misc hosts we do not want alerts on ##HOST=mac1.hmgc.mcw.edu ## IGNORE ##HOST=canadeo-disk.brc.mcw.edu ## IGNORE ##HOST=sharpe.phys.mcw.edu ## IGNORE HOST=bc1s7.phys.mcw.edu IGNORE # # Now on to our regular monitoring rules for things we DO want alerts on # HOST=uram.hmgc.mcw.edu MAIL brodie at phys.mcw.edu REPEAT=30 RECOVERED FORMAT=TEXT MAIL kent.brodie at pager.mcw.edu REPEAT=30 RECOVERED FORMAT=SMS ##SERVICE=disk ## MAIL brodie at phys.mcw.edu REPEAT=30 RECOVERED FORMAT=TEXT HOST=loki.phys.mcw.edu SERVICE=disk MAIL brodie at phys.mcw.edu REPEAT=60 RECOVERED FORMAT=TEXT SERVICE=conn MAIL sysadmins REPEAT=30 RECOVERED FORMAT=TEXT MAIL pagers REPEAT=30 RECOVERED FORMAT=SMS SERVICE=oracle MAIL szacher at hmgc.mcw.edu REPEAT=30 RECOVERED FORMAT=TEXT MAIL sysadmins REPEAT=30 RECOVERED FORMAT=TEXT MAIL kent.brodie at pager.mcw.edu REPEAT=30 RECOVERED FORMAT=SMS [hobbit at coffman etc]$

2 1