- Xymon - Xymon mailman web

Extra value in the memory graph
by Bart_Gillis＠msp.be 21 Nov '06

21 Nov '06

Hello, I wrote a script that monitors the committed memory usage of a windows server. This script uses BBWin 0.8 client to load these values to hobbit. The script creates a file "memcommit" in the temp file of the BBWin client. Each 30 min the values ("realmempct") are send to the Hobbit where an extra column appears on the display server. I modified the hobbitserver.cfg, TEST2RRD="cpu=la,disk,... ,memcommit=ncv" NCV_memcommit="realmempct:GAUGE" And guess... A file memcommit.rrd was created that designed a new graph under the value of the new test. So far, all went fine. However, I'm not interested in 3 or 4 separate columns for my extra memory tests though I'd like to add this extra value in the existing memory graph. I have one restriction; my BBWin client does not allow files that contain a dot. (ex: memory.memcommit. These are not sent to the Hobbit server). In other words, is there a way that the Hobbit server can create a memory.memcommit.rdd file instead of the memcommit.rdd and how can I avoid that the extra column memcommit is created? Thanks in advance, Bart

2 1

Mixed disk graph results
by rmunsch＠solutionsforprogress.com 21 Nov '06

21 Nov '06

Hello list, I need to troubleshoot my disk usage graphs. Running debian, i have both values and NaN on the same graph, when viewed under trends or Disk itself. It's odd. Frex, i'm looking now at a graph that has values for / and /boot, and lines are plotted; but the same graph is summarizing nan for /home and /shared. All other tests are working fine. It's only disk usage, across multiple hosts, that remains wonky. -- Rob Munsch Solutions For Progress IT www.solutionsforprogress.com

1 0

bbtest error
by lebarber＠gmail.com 21 Nov '06

21 Nov '06

I am occasionally seeing the following error report from bbtest: IO error in SSL_connect to https (47873/tcp) on host x.y.z.a: error:00000000:lib(0):func(0):reason(0) Does anybody know what this signifies? (my apologies for the duplicate post) Thanks, Larry Barber

3 5

theshire and SF
by gjohnson＠trantor.org 21 Nov '06

21 Nov '06

Ok, I'm getting really frustrated with SF's limitations. For a project like The Shire I'm beginning to believe that it needs something a bit more flexible. I've decided to use a wiki to hold the site but most wikis require write privs to the filesystem...if only for the sessions. SF mounts all sites as read-only which really limits what can be done (hell, I was playing around with adding a page counter and couldn't even get that to work properly). I'm currently looking at a couple of other options and was hoping to have something up this week while I was off. =G=

2 2

[hobbit] client-local.cfg question
by cdanger2k＠yahoo.com 20 Nov '06

20 Nov '06

Thanks, I appreciate the heads up! ----- Original Message ---- From: Charles Jones <jonescr at cisco.com> To: hobbit at hswn.dk Sent: Monday, November 20, 2006 12:37:49 PM Subject: Re: [hobbit] client-local.cfg question Yes. See the CLASS option in the bb-hosts man page. Basically you can define a CLASS (group) of hosts, and then in client-local.cfg you can do: [classname] log:`find /var/log/httpd -type f -maxdepth 1` size<10000000 -Charles Chuck Danger wrote: Thought I would try this again, maybe I caught people on a weekend... ----- Original Message ---- From: Chuck Danger <cdanger2k at yahoo.com> To: hobbit at hswn.dk Sent: Friday, November 17, 2006 4:25:31 PM Subject: [hobbit] client-local.cfg question Is there a way in client-local.cfg to group hosts by "like" hostnames? For example, if I have: web01 web02 web03 web04 web05 And I want to setup file monitoring in /var/log/httpd, is there some way to group them so that I don't have to define each one? I'd like to do something like: [web*] log:`find /var/log/httpd -type f -maxdepth 1` size<10000000 instead of having to do: [web01] ... [web02] ... [web03] ... Thanks- Jay

1 0

[hobbit] client-local.cfg question
by cdanger2k＠yahoo.com 20 Nov '06

20 Nov '06

Thought I would try this again, maybe I caught people on a weekend... ----- Original Message ---- From: Chuck Danger <cdanger2k at yahoo.com> To: hobbit at hswn.dk Sent: Friday, November 17, 2006 4:25:31 PM Subject: [hobbit] client-local.cfg question Is there a way in client-local.cfg to group hosts by "like" hostnames? For example, if I have: web01 web02 web03 web04 web05 And I want to setup file monitoring in /var/log/httpd, is there some way to group them so that I don't have to define each one? I'd like to do something like: [web*] log:`find /var/log/httpd -type f -maxdepth 1` size<10000000 instead of having to do: [web01] ... [web02] ... [web03] ... Thanks- Jay

2 1

Enable/Disable by page name instead of hostname(s)?
by jonescr＠cisco.com 20 Nov '06

20 Nov '06

Is this possible to do? Heres the situation I am trying to avoid. When there is going to be maintenance activities on all of the servers (let's say a kernel upgrade). Putting them all into disable mode literally floods the alert pagers and email with hundreds of messages (they have NOTICE tag). I have my Hobbit pages built thus that the main page has subpages for Prod, Stage, and Dev. Each one of those has sub-pages for Web, App, and Database tiers. It would be nice if there was a way to enable/disable/acknowledge by page name, or a pre-defined groupname. Then for example one could disable the "Prod" group, and only one disable notice alert gets sent, instead of one for every host. Alternatively, perhaps an option could be added to allow a host(s) to be "silently" enabled/disabled, such that no alerts are sent to addresses using the NOTICE tag. Other ways I can think of to deal with this: 1. Do not use NOTICE tag. Not really an option, but could temporarily remove it before doing an enable/disable of many hosts in order to reduce the spam. 2. Use DOWNTIME in bb-hosts, instead of using the interactive enable/disable. This would work but requires manual editing of the bb-hosts file, which takes away the usefulness of enabling/disabling from the web interface. Both of the above provide a workaround, but are not optimal. -Charles

2 1

[patch] debian hobbit client init script
by gn1＠sanger.ac.uk 20 Nov '06

20 Nov '06

Hi, As I said earlier we want the hobbit client script to remove the temporary files and kill the rogue processes when stopped, mainly so the hobbit-client installation/upgrade process through cfengine runs smoothly. Here is a patch to the debian hobbit client init script to do the following actions when the hobbit client is stopped: - remove the /var/run/hobbitclient-runtime.cfg file - kill the unwanted vmstat processes that were left running - remove any files owned by hobbit from /tmp Patch may be applied on 4.2.0 with allinone patch applied, using: cd hobbit-4.2.0 patch -p1 < /path/to/patch/hobbitclient_clean.patch Comments to help polishing this solution are more than welcome! Cheers Gildas

2 2

Hobbit
by pretocj2＠staff.intekom.com 20 Nov '06

20 Nov '06

Hi Im am a sysadmin fo a ISP tipe environment. We have been using Nagios as our monitoring tool. I do admit that I like Nagios and prefer Nagios over using the Hobbit. Unfortuanately some people with too much say in our organisation decided to use Hobbit. These people are big hobbit fans and I even found them on mailing lists trying to convince Nagios users to switch to hobbit. Since one of my main functions is maintaining the monitoring environment, this decision was a bit of a let down. Hobbit is easier to configure, I give you that. We are running Nagios and Hobbit side by side and obviously when new servers are build nagios is left out and the kickstarts are built to install hobbit. But I could do some comparisons and I started collecting data that makes hobbit look not so reliable as Nagios. Nagios also has some functionality that I could not find in hobbit. I have never been on any hobbit mailing list. Since the decision was made, and that is now the road we are going. I would like to see some changes/additions, or make some suggestions to the hobbit. Where can I send my suggestions to? I am currently a bit anti hobbit, for good reason, but that could change, I hope. Regards CJ

3 2

no mailq graph?
by stefan.freisler＠aspect-online.de 20 Nov '06

20 Nov '06

Hi, i don't get an mailq graph. What is to do? rrd file will not created I use Hobbit 4.2.0 Part of my hobbitserver.cfg TEST2RRD="cpu=la,disk,inode,qtree,$PINGCOLUMN=tcp,http=tcp,dns=tcp,dig=tcp,time=ntpstat,vmstat,iostat,netstat,temperature,apache,apache1=apache,ntp=clock,bind,sendmail,mailq,nmailq=mailq,socks,bea,iishealth,citrix,bbgen,bbtest,bbproxy,hobbitd,files,procs=processes,ports,clock,lines,temp,oracle,sar-io,sar-rpca,sar-rqlla,memory,sar-cpu,sar-cswitch,sar-kernel,r-huk24-a,r-huk24-b,r-huk24-c,r-div,r-allianz,r-axakfz,r-directline,r-ineas,r-sdirekt,r-ineas,spam" # This defines which RRD files to include on the "trends" column webpage, # and the order in which they appear. GRAPHS="la,disk,inode,qtree,files,processes,memory,users,vmstat,iostat,tcp.http,tcp,ncv,netstat,ifstat,mrtg::1,ports,temperature,ntpstat,apache,bind,sendmail,mailq,socks,bea,iishealth,citrix,bbgen,bbtest,bbproxy,hobbitd,clock,lines,temp,oracle,sar-io,sar-rpca,sar-rqlla,sar-memory,sar-cpu,sar-cswitch,sar-kernel,r-directline,r-huk24,r-huk24b,r-huk24c,r-div,r-allianz,r-axakfz,r-directline,r-ineas,r-sdirekt,response,spam" thx for Help Stefan

1 0