June 2007 - Xymon - Xymon mailman web

Multiple display servers
by sigurdur＠raforninn.is 22 Jun '07

22 Jun '07

Hi. I'm having a little trouble with configure'ing multiple hobbit servers to communicate properly between themselves. I have one main central server which gets all information from the other servers and doesnt need to ping anything or check anything, just display the results that the other servers have. The other servers check on their own private networks how the computer statuses are (the hosts send their information to the hobbit) and then the hobbit relays the information to the central server. The problem is that with some (and for a strange reason not others) hobbit relays seem to indicate that every computer is down for a couple of minutes (usually around 2) and then up for 3 minutes. When I check on the relay server, every conn check is available for many days, however on the central server the log is filled with disconn's. Now, how do I fix this? Hobbit 4.2.0, Ubuntu. With regards, Sigurður Guðbrandsson

1 0

Hobbit windows client: configuration
by yf.liang＠yahoo.ca 22 Jun '07

22 Jun '07

I've been using BBWin08.msi for a while, and would like to see CPU usage reported in percentage, like on Unix's, instead of load numbers. Does 0.9 allow that? Or would I have to customize the source code and compile it myself? Thanks, Ye-Fee --------------------------------- All new Yahoo! Mail - --------------------------------- Get a sneak peak at messages with a handy reading pane.

3 2

Installing perl-SNMP_Session
by FRamaekers＠ailife.com 21 Jun '07

21 Jun '07

How does one install this? All I got was a tarball and very little installation instructions. This package is a requirement for 'devmon'. TIA, Frank M. Ramaekers Jr. Systems Programmer MCP, MCP+I, MCSE & RHCE American Income Life Insurance Co. Phone: (254)761-6649 1200 Wooded Acres Dr. Fax: (254)741-5777 Waco, Texas 76701 ---------------------------------------- This message contains information which is privileged and confidential and is solely for the use of the intended recipient. If you are not the intended recipient, be aware that any review, disclosure, copying, distribution, or use of the contents of this message is strictly prohibited. If you have received this in error, please destroy it immediately and notify us at PrivacyAct at ailife.com. ---------------------------------------- This message contains information which is privileged and confidential and is solely for the use of the intended recipient. If you are not the intended recipient, be aware that any review, disclosure, copying, distribution, or use of the contents of this message is strictly prohibited. If you have received this in error, please destroy it immediately and notify us at PrivacyAct at ailife.com.

3 4

NCV RRD handler that will create multiple rrd files from a single test
by tomg＠mcclatchyinteractive.com 21 Jun '07

21 Jun '07

This topic has come up on the list a couple of times, but I don't think I've ever heard a definitive answer. I really need the ability for the NCV module (or a different module) to create multiple rrd files using data from a single test, just like the way the do_disk rrd handler works. Over the last month I have been working to get a lot of my networking equipment monitored and graphed using devmon, and there are many situations (such as concurrent sessions per virtual server on my load balancer) where the addition of a new monitored parameter (i.e adding a new virtual server) to the exisiting test will cause havoc for the ncv rrd handler. Is there any chance that this kind of functionality will make it into Hobbit? It'd be cool to have an NCV module that can handle both cases, or an additional NCV module to produce multiple rrd files and works alongside the traditional one. Tom -- Tom Georgoulias Systems Engineer McClatchy Interactive tomg at mcclatchyinteractive.com

6 7

Why BBSVCNAME,BBHOSTSVC invalid in my alert script?
by casedj＠21cn.com 21 Jun '07

21 Jun '07

Hobbit 4.2.0 on RHEL4, BB client 1.08d on Windows XP cici. hobbit-alerts.cfg: HOST=cici SERVICE=procs,svcs SCRIPT /home/hobbit/server/etc/alert-svcs.sh cici Some environment variables,e.g,BBSVCNAME,BBHOSTSVC are used in alert script, but they actually weren't shown as specific service or process that triggered the alert while the service or process reached down, just generic string procs. alert-svcs.sh #!/bin/sh echo "$BBHOSTSVC reached error state" |write hobbit echo "$BBSVCNAME reached error state" |write hobbit The result displayed as below: Message from hobbit at my_linux on <no tty> at 22:21 ... cici.procs reached error state EOF Message from hobbit at my_linux on <no tty> at 22:21 ... procs reached error state EOF Thanks a lot. Regards, Case

3 5

[hobbit] Well, duh! I didn't know Hobbit/rrdtool did that!
by Thomas.Kern＠hq.doe.gov 21 Jun '07

21 Jun '07

Yeah, that is great and you can do more with the time scale and size in the URL. With less than a minute of editing the URL, I can print off a monthly CPU chart that I used to spend an hour creating in Windows Excel. Thomas Kern 301-903-2211 ----- Original Message ----- From: Haertig, David F (Dave) <haertig at avaya.com> To: hobbit at hswn.dk <hobbit at hswn.dk> Sent: Sun Jun 03 18:47:51 2007 Subject: [hobbit] Well, duh! I didn't know Hobbit/rrdtool did that! After messing around and messing around trying to set graph timeframes and vertical axis limits, I just had a duh! moment. You don't need that. Those little magnifying glass icons? Yeah, I had clicked on them before expecting to zoom in on the graph. But it never zoomed in ... it just showed the same graph again on a seperate webpage. I figured it was some broken feature. My duh! moment just came when I accidently drug my mouse cursor over part of the "non zoomed" graph and noticed, "Hey! I wonder why it highlighted only that small portion of the graph?!" Release the mouse button and it becomes very evident why. DUH!!! How is it possible for me to have been so stupid?! :-o

2 2

[bug] bb2.html show red:nopropagate:
by charles.goyard＠orange-ftgroup.com 21 Jun '07

21 Jun '07

Hi, I encounter what looks much like a bug in bbgen. I have two hosts in failover, so I have the following logic : cwd-resa01o.cwd93.cvf.is_master = green cwd-resa02o.cwd93.cvf.is_master = red host-cluster = green (host1.is_master XOR host2.is_master with bbcombogen) So, either host1 or host2 is red all the time, and I don't want to propagate it. So I did : 10.193.173.21 cwd-resa01o.cwd93.cvf # ssh NOPROPRED:is_master,procs_master 10.193.173.22 cwd-resa02o.cwd93.cvf # ssh NOPROPRED:is_master,procs_master When cwd-resa02o.cwd93.cvf.is_master gets red, it does not show on bb2, but when cwd-resa01o.cwd93.cvf.is_master, it propagates and I get the following on bb2.html : is_master cwd-resa01o.cwd93.cvf is_master:red:nopropagate:2m It's somewhat strange to see "nopropagate" on the bb2 page... The debug log shows : load_bbhosts: -- got line '10.193.173.21 cwd-resa01o.cwd93.cvf # ssh NOPROPRED:is_master,procs_master' init_state(cwd-resa01o,cwd93,cvf.is_master, 135215936, ...) find_or_create_column(is_master) init_state: hostname=cwd-resa01o.cwd93.cvf, testname=is_master, color=5, acked=0, age=2m, oldage=0, propagate=0, alert=0 init_state(cwd-resa02o,cwd93,cvf.is_master, 135264320, ...) find_or_create_column(is_master) init_state: hostname=cwd-resa02o.cwd93.cvf, testname=is_master, color=5, acked=0, age=33s, oldage=0, propagate=0, alert=0 Host: cwd-resa02o.cwd93.cvf, ip: 10.193.173.22, name: cwd-resa02o.cwd93.cvf, color: 0, old: 0, anywaps: 0, wapcolor: -1, pretitle: '(NULL)', noprop-y: (NULL), noprop-r: ,is_master,procs_master,, noprop-p: (NULL), noprop-ack: (NULL), waps: (NULL) Test: is_master, alert 0, propagate 0, state 5, age: 33s, oldage: 0 Host: cwd-resa01o.cwd93.cvf, ip: 10.193.173.21, name: cwd-resa01o.cwd93.cvf, color: 4, old: 0, anywaps: 0, wapcolor: -1, pretitle: '(NULL)', noprop-y: (NULL), noprop-r: ,is_master,procs_master,, noprop-p: (NULL), noprop-ack: (NULL), waps: (NULL) Test: is_master, alert 0, propagate 0, state 5, age: 2m, oldage: 0 I run bbgen as: bbgen --recentgifs --subpagecolumns=2 --report --nopropred=procs_master,is_master I also tried without NOPROPRED and only --nopropred, same thing. May I say that I ran bbgen against a minimalistic bb-hosts file, with only these two hosts. Maybe that one already had been worked out ? -- Charles Goyard - charles.goyard at orange-ftgroup.com - (+33) 1 45 38 01 31 Orange Business Services - online multimedia // ingénierie

1 0

Local test limits
by Bart.Gillis＠msp.be 21 Jun '07

21 Jun '07

Hello, In the bb-hosts file you can set limits on the network test by specifying the badTEST x:y:z parameter. Is there a way that you can specify such limits for local tests (cpu, custom tests,...) to? If I set, for instance badcpu 1:2:3 as a parameter, this is not displayed in the configuration report and does not work: Network tests Service NK C/Y/R limits Specifics conn No 1/2/3 Local tests Service NK C/Y/R limits Configuration (NB: Thresholds on client may differ) cpu No -/-/- UNIX - Yellow: Load average > 1.5, Red: Load average > 3.0 Windows - Yellow: CPU utilisation > 80%, Red: CPU utilisation > 95% Best Regards Bart Gillis

2 1

Local test limits
by Bart.Gillis＠msp.be 21 Jun '07

21 Jun '07

Hello, In the bb-hosts file you can set limits on the network test by specifying the noTEST x:y:z parameter. Is there a way that you can specify such limits for local tests (cpu, custom tests,...) to? If I set, for instance nocpu 1:2:3 as a parameter, this is not displayed in the configuration report and does not work: Network tests Service NK C/Y/R limits Specifics conn No 1/2/3 Local tests Service NK C/Y/R limits Configuration (NB: Thresholds on client may differ) cpu No -/-/- UNIX - Yellow: Load average > 1.5, Red: Load average > 3.0 Windows - Yellow: CPU utilisation > 80%, Red: CPU utilisation > 95% Best Regards Bart Gillis

1 0

MAXMSG_*
by dbourque＠weatherdata.com 20 Jun '07

20 Jun '07

Hi, I set these values in the hobbitserver.cfg: MAXMSG_STATUS="1024" MAXMSG_CLIENT="1024" MAXMSG_DATA="1024" I still get some : Latest errormessages: Oversize data/client msg from n.n.n.n truncated (n=2003124, limit 1048576) If I got the client info dump on the web page for that client. It's always back to green when I go look at it. When it's green, if I cut and paste all the text into a text file and and run du -k on it, I get 80. How can that output possibly grow over twenty times to trigger an hobbitd warning ??? Am I not understanding exactly what those limits apply to ? Thanks! -Dan

1 1