May 2008 - Xymon - Xymon mailman web

[hobbit] hobbit client
by forums＠triadbiz.rr.com 05 May '08

05 May '08

I am alerting on yellow and red but shouldn't the page still go yellow or red in the hobbit display even if I wasn't. Josh From: Galen Johnson [mailto:Galen.Johnson at sas.com] Sent: Monday, May 05, 2008 11:36 AM To: hobbit at hswn.dk Subject: RE: [hobbit] hobbit client Unless you have alert on yellow enabled, it won't alert until disk reaches 95% unless I'm missing something. =G= From: Joshua Krause [mailto:forums at triadbiz.rr.com] Sent: Monday, May 05, 2008 11:29 AM To: hobbit at hswn.dk Subject: [hobbit] hobbit client I've got a group of servers running the hobbit client and my disk is not being alerted on. Here is a sample from the disk page in hobbit: Filesystem Size Used Avail Use% Mounted on /dev/sda2 7.9G 6.8G 749M 91% / /dev/sda1 99M 34M 60M 37% /boot /dev/sda3 24G 429M 22G 2% /export/home Status unchanged in 58 days,22 hours, 27 minutes Status message received from 172.31.129.185 Client <http://ops1/monitor-cgi/bb-hostsvc.sh?CLIENT=gsonc-spg-vod-t16-4> data available <http://ops1/monitor-cgi/hobbitgraph.sh?host=gsonc-spg-vod-t16-4&service=dis k&graph_width=576&graph_height=120&disp=gsonc%2dspg%2dvod%2dt16%2d4&nostale& color=green&action=menu> And here is what I have in the /opt/hobbit/server/etc/hobbit-clients.cfg file: HOST=%gsonc-spg-ops-1 LOAD 9.0 15.0 PROC snmpd PROC snmptt PROC sec.pl PROC snmptrapd PROC mysqld HOST=%gsonc-spg-ops-2 PROC radiusd PROC mysqld HOST=%gsonc-spg-ops-3 PROC syslog-ng PROC mysqld HOST=%gsonc-spg-omea-1 PROC cron LOAD 7.0 10.0 HOST=%gsonc-spg-sunbox LOAD 11.5 12.5 MEMSWAP 85 95 HOST=%gsonc-spg-dncs FILE /export/home/dncs/mystro/bfs/Compass_OOB/rta.grp.z SIZE>1200 TRACK=rta LOAD 10.0 15.0 DISK / 95 98 DISK /cdrom IGNORE HOST=%gsonc-spg-mas* FILE /usr/local/mystro/conf/rtareports_headendinfo.properties MTIME=1128514608 TRACK=rta-reports MEMPHYS 100 101 MEMSWAP 75 90 MEMACT 97 98 HOST=%gsonc-spg-rtc* LOAD 17.0 22.0 HOST=%gsonc-spg-vod-t* LOAD 17.0 22.0 HOST=%gsonc-spg-vodgs* LOAD 17.0 22.0 PROC cron DEFAULT # These are the built-in defaults. UP 1h LOAD 5.0 10.0 DISK * 90 95 MEMPHYS 100 101 MEMSWAP 75 90 MEMACT 90 97 I highlighted in bold the group that deals with one of the servers in question. I should use the DEFAULT group for everything except load. And load is working right and the uptime reports back good also. So it is showing the alerting on uptime via the DEFAULTS and it is alerting on LOAD with the exception. So why is DISK not alerting out or even changing colors. I have tried these to variations but they didn't work either: HOST=%gsonc-spg-vod-t* LOAD 17.0 22.0 DISK / 90 95 HOST=%gsonc-spg-vod-t* LOAD 17.0 22.0 DISK * 90 95 And neither of those worked. Any help greatly appreciated. Thanks, Josh

1 0

purple server side test?
by michael.nemeth＠lmco.com 05 May '08

05 May '08

Below is a server script I adapted for showing a back-up (save prosses) is running. I have a trigger set up for the proc (yellow if the save string is in the ps)column that works, but I wanted separate column. Sometimes I get purples with below script when backups are running and confirmed by the proc column but mostly it works, How can I be getting purples? #!/usr/bin/perl #*----------------------------------------------------------------------------*/ #* Hobbit client message processor. */ #* */ #* This perl program shows how to create a server-side module using the */ #* data sent by the Hobbit clients. This program is fed data from the */ #* Hobbit "client" channel via the hobbitd_channel program; each client */ #* message is processed by looking at the [who] section and generates */ #* a "login" status that goes red when an active "root" login is found. */ #* */ #* Written 2007-Jan-28 by Henrik Storner <henrik at hswn.dk> */ #* */ #* This program is in the public domain, and may be used freely for */ #* creating your own Hobbit server-side modules. */ #* */ #*----------------------------------------------------------------------------*/ # $Id: rootlogin.pl,v 1.1 2007/01/28 12:42:34 henrik Exp $ my $bb; my $bbdisp; my $hobbitcolumn = "bkup"; my $hostname = ""; my $msgtxt = ""; my %sections = (); my $cursection = ""; sub processmessage; # Get the BB and BBDISP environment settings. $bb = $ENV{"BB"} || die "BB not defined"; $bbdisp = $ENV{"BBDISP"} || die "BBDISP not defined"; # Main routine. # # This reads client messages from <STDIN>, looking for the # delimiters that separate each message, and also looking for the # section markers that delimit each part of the client message. # When a message is complete, the processmessage() subroutine # is invoked. $msgtxt contains the complete message, and the # %sections hash contains the individual sections of the client # message. while ($line = <STDIN>) { if ($line =~ /^\@\@client\#/) { # It's the start of a new client message - the header looks like this: # @@client#830759/HOSTNAME|1169985951.340108|10.60.65.152|HOSTNAME|sunos|sunos # Grab the hostname field from the header @hdrfields = split(/\|/, $line); $hostname = $hdrfields[3]; # Clear the variables we use to store the message in $msgtxt = ""; %sections = (); } elsif ($line =~ /^\@\@/) { # End of a message. Do something with it. processmessage(); } elsif ($line =~ /^\[(.+)\]/) { # Start of new message section. $cursection = $1; $sections{ $cursection } = "\n"; } else { # Add another line to the entire message text variable, # and the the current section. $msgtxt = $msgtxt . $line; $sections{ $cursection } = $sections{ $cursection } . $line; } } # we watch the [who] section of the client message and alert # if there is a root login active. sub processmessage { my $color; my $summary; my $statusmsg; my $cmd; my $cmdl; # Dont do anything unless we have the "ps" section return unless ( $sections{"ps"} ); # Is there a "root" login somewhere in the "who" section? # Note that we must match with /m because there are multiple # lines in the [who] section. if ( ($cmdl) = $sections{"ps"} =~ /(save -s.*)/ ) { $color = "yellow"; $summary = "There MAYBE active backups" ; $statusmsg = "&yellow " . $cmdl . $sections{"ps"}; } else { $color = "green"; $summary = "Ok"; $statusmsg = "&green No backup active\n\n"; } # Build the command we use to send a status to the Hobbit daemon $cmd = $bb . " " . $bbdisp . " \"status " . $hostname . "." . $hobbitcolumn . " " . $color . " " . $summary . "\n\n" . $statusmsg . "\""; # And send the message system $cmd; }

1 0

Hobbit to IPPhone
by jonescr＠cisco.com 05 May '08

05 May '08

I just cobbled together a script that puts a hobbit display on a Cisco IPPhone (7960). The display is pretty spartan since we only have half a dozen lines to put text on :) The script is run out of cron on my linux workstation, and requires access to an http server, and knowing your phones login and password (same as your user login to call manager). If anyone is interested in this script I will post it here or send it to the Shire. I just did it mainly for fun, but with some effort you could easily even acknowledge alerts right from your phone :) -Charles

9 37

Color doesn't change by alert ?!
by martin.flemming＠desy.de 05 May '08

05 May '08

Hi ! I've got on my main page the section (page) K5cron ~/etc/bb-hosts page K5cron K5cron group-only conn|cpu|cpu2|disk|ssh|k5jobs|info|memory|msgs|procs|trends2|trends|TRENDS:*,vmstat:vmstat|vmstat3|vmstat4|vmstat5,netstat:netstat1|netstat2|netstat3 ,tcp:conn 131.169.56.53 k5cron01 # ssh k5jobs 131.169.56.53 k5cron02 # ssh The color of the icon of K5cron is green But the machine k5cron02 is offline, Conn is RED, all other white for k5cron02, if i go to the page K5cron ... I've got alert-mails, but the icon on the main-page for k5cron stay green ! ... why ? thanks & cheers Martin ______________________________________________________ Martin Flemming DESY / IT office : Building 2b / 008a Notkestr. 85 phone : 040 - 8998 - 4667 22603 Hamburg mail : martin.flemming at desy.de ______________________________________________________

1 0

Checkpoint High Availability Monitoring
by linuxmasterjedi＠free.fr 05 May '08

05 May '08

Hi, once again ;) I would like to monitor the HA cluster between 2 Checkpoint Firewall. Several time a month, the cluster failed down, sound like one of the numerous Ethernet card seems to be off during a couple of second, thus the HA cluster switch to safe mode. I would like to detect this nasty state from the Hobbit server to broadcast an alarm. I don't know how the 2 Firewalls exchange health information (maybe via Heartbeat). Anyone already figure out how to monitor this issue? Thanks by advance, LMJ

4 10

Depends feature
by David.Stuffle＠adesa.com 02 May '08

02 May '08

I'm using the "depends" feature heavily for HTTP tests and it is doing some weird things. When a HTTP test fails, the status turns clear saying the depended on test has failed, but it has not failed. But then the status turns red, then clear, then red, then clear... See the history below. Each clear status increases by 5 seconds. And I am positive that the depended on test has not failed. Mon Apr 21 20:24:37 2008 0:04:36 Mon Apr 21 20:24:11 2008 0:00:26 Mon Apr 21 20:19:30 2008 0:04:41 Mon Apr 21 20:19:09 2008 0:00:21 Mon Apr 21 20:14:23 2008 0:04:46 Mon Apr 21 20:14:07 2008 0:00:16 Mon Apr 21 20:09:16 2008 0:04:51 Mon Apr 21 20:09:06 2008 0:00:10 Mon Apr 21 20:04:09 2008 0:04:57 Mon Apr 21 20:04:04 2008 0:00:05 Mon Apr 21 20:00:04 2008 0:04:00 Syntax is like this: 0.0.0.0 http-a-chryslerfirst # noconn http://www.chryslerfirst.com=192.168.254.94 depends=(http:was-b-chryslerfirst/http,was-c-chryslerfirst/http) Anyone experience the same thing or have any idea how to troubleshoot? Using 20080103 snapshot. Thanks! D. Stuffle

2 1

Problem using spaces in hobbit-clients.cfg (SVCs)?
by Matthew.Roberts＠nt.gov.au 01 May '08

01 May '08

Hi All, I am having the same problem as the one below except I am using hobbit 4.2 with the allinone patch and the bbwin patch. I was unable to find a solution within the Mailing List archives so does anyone have a solution? Or is it something which is an ongoing issue? Regards, Problem using spaces in hobbit-clients.cfg (SVCs)? To: hobbit (at) hswn.dk Subject: Problem using spaces in hobbit-clients.cfg (SVCs)? From: Magnus Carlebjörk <magnus (at) carlebjork.se> Date: Fri, 4 Apr 2008 11:06:54 +0200 (CEST) References: <86690B47898A454CACD55248748A527257888C (at) ac-hq-exch2.corp.americashloans.net> <OF1EB9A849.A1A7C19F-ON86257420.0059CC72-86257420.0059FDA2 (at) mck.us.ray.com> <68e737a10804040126s5aac601bk331c22e7511d0432 (at) mail.gmail.com> Hi, I'm having problems using spaces in service names in hobbit-clients.cfg. Here is an example block for a host: HOST=ABCD123 UP 1h SVC "Pml Driver HPZ12" startup=automatic status=started color=red PROC pageant.exe 1 1 LOAD 80 95 DISK * 90 95 MEMPHYS 90 101 MEMSWAP 80 90 MEMACT 90 97 Even though the service is started, hobbit cannot find the service: red Fri Apr 04 10:57:22 2008 - Services NOT ok &red Pml Driver HPZ12 is Not Found req automatic started However, looking at the output on the status page, I can see that the service is reported and started: Name StartupType Status DisplayName Pml Driver HPZ12 automatic started Pml Driver HPZ12 I'm using 4.3.0-0.20080402. String handling bug? Regards, -- Magnus Carlebjork Stockholm, Sweden +46 76 116 9008

1 0