October 2008 - Xymon - Xymon mailman web

White-clear dots where they shouldn't be
by eric.jacobs＠thomastechsolutions.com 09 Oct '08

09 Oct '08

Lately, whenever I install a new Linux client (hobbit 4.20 with allinone patch - server the same), I get white dots (which eventually go to clear dots) for "files", "msgs", "ports" and "procs". Usually I have procs defined in hobbit-clients.cfg and I change permissions on /var/log/messages. Hobbitd-client on server is running with "--no-clear-files --no-clear-ports --no-clear-msgs". I also don't get the "client data" link. This didn't use to happen. So, I'm confused. Anyone else get this? Eric Jacobs

3 2

Additional reports
by graham.nayler＠hallmarq.net 08 Oct '08

08 Oct '08

Hi, We are monitoring a certain type of activity on our client machines using the external script mechanism. I've got it set up with the usual status flags and graphs etc. and that's working fine. There're now two things I need to add: 1) Numerical display of the values displayed in the graphs. Obviously these could be obtained from RRDtool, but is there a way of displaying them e.g. on a subpage from the webpages? 2) We want to display a list of the top N clients reporting on a particular metric, say on either "Weekly" or "Hourly" metrics under an "Activity" column. It looks like this could be obtained from the "Top-Changes Report" menu option, but is there any way of canning the form fields and plugging it into the menu system (or anywhere else it would be appropriate) Thanks, Graham Nayler Hallmarq: Standing Equine MRI http://www.hallmarq.net Hallmarq Veterinary Imaging Ltd, Unit 5, Bridge Park , Merrow Lane, Guildford, GU4 7BF UK Registered in England and Wales No. 4061791 DISCLAIMER: This email and any files transmitted with it are confidential to the intended recipient(s). If you have received this email in error please notify the author by replying to this email and deleting it and any copies from your system. Any views or opinions contained in this email are those of the author and not necessarily endorsed by Hallmarq.

1 1

Disk graphs intermittantly missing disk info (resent in plaintext)
by TRBrand＠cvs.com 08 Oct '08

08 Oct '08

[Apparently there was a problem the first time I sent this... trying again, text only...] Problem: Disk Utilization graphs are not showing the correct information when one partition is not mounted. Disk utilization trend graphs are not consistant in displaying the information. I am monitoring 3 file systems: '/', '/data, and '/backup'. The /backup partition is only mounted while backups are actually running which only takes a few minutes. When the backup completes, the /backup drive is unmounted. In all cases, the text (from 'df') on the web page correctly lists the partitions that are mounted and correctly displays any disk space alerts: -------------------- /data (27% used) has reached the WARNING level (25%) Filesystem 1024-blocks Used Available Capacity Mounted on LABEL=ROOT1 8652060 3039788 5172772 38% / /dev/sda7 14740312 3753364 10238176 27% /data -------------------- When the backup partition is mounted, the disk utilization image displays all three partitions in the line graphs and lists all 3 backup' with the cur/max/min/avg data below the line graph. When the backup partition is not mounted, the image displays only the '/' root and '/backup' partition in the line graph, and only lists '/' and '/backup' with the cur/max/min/avg data below the line graph. The '/data' partition is not displayed/listed anywhere in the image. Looking for some suggestions as to why / how this is happening and if there is something I can do to fix this. thanks for any suggestions/pointers, Tom

1 0

mpstat and zonestat monitors
by annaj＠hi.is 08 Oct '08

08 Oct '08

Hi all. Solaris 10 has a virtualization capability which is the ability to create so called zones, that effectively act like independent machines. In The Shire are mpstat and zonestat monitors for monitoring zones on Solaris 10. I would like to know if it is possible to monitor all zones on a server so that a client on each virtual machine is not necessary. I do not like the alternative, which is to install a hobbit client on each virtual machine. PS. Henrik: It is good to know You are in good health. -- Kindest Regards, Anna Jonna Ármannsdóttir, %& A: Because people read from top to bottom. Unix System Aministration, Computing Services, %& Q: Why is top posting bad? University of Iceland.

10 15

ERROR: Value too large to be stored in data type
by James.Roberts＠hants.gov.uk 08 Oct '08

08 Oct '08

I am trying to chart some very big files in rrd. However I am getting this under the files check? [file:/export/nim/mksysbs/sapr3s.mksysb] ERROR: Value too large to be stored in data type Is there anyway round this or will I need to write a custom check? James Roberts IT Consultant - IT Services Hampshire County Council Room 415,Elizabeth II Court, The Castle, Winchester, Hants SO23 8US Internal HPSN: (200) 5037 Tel: +44 (0)1962 845037 Fax: +44 (0)1962 847434 james.roberts at hants.gov.uk Web: WWW.Hants.Gov.UK

1 0

Questions about centralised configuration of hobbit clients
by oyvind.bjorge＠telenor.com 08 Oct '08

08 Oct '08

Hi all! We are considering migrating from BigBrother to Hobbit. We have been very satisfied with BB, but after reading about hobbit, it seems like a big improvement. Also it looks like BB is a dead end, so I hope for Hobbit to be the future, still being developed. I have some questions relating too the centralised configuration of the clients. How is this performed? I have read about how it is configured, but was wondering about how this is transferred to the client. Is it the same port that is used? Is this checked every time the test is running? Is it possible to use this mechanism in ext-scripts, by adding more parameters to the centralised config-file? If so, how is this addressed from the ext-scripts? I also have read something about possibility of upgrading clients from remote. How is this performed? If this is pushed from the server, I guess it is needed with some opening from the server towards the client (not only from the client to the server as for BB). Or is the client checking for new updates on a regular basis? Is this mechanism used also for distributing ext-scripts to clients? In my BB installation today I have created my own system for distributing config-files and ext-scripts to client, for having a centralised handling of this. This is both the standard bb-config files and one separate file for my ext-scripts, and I created a website for editing these files so it would be easier for other users to edit this config and distributing the files. If hobbit have built-in (or add-on) functionality for this, so that my solution for this would become obsolete, it would be great. Hope for someone to clearify these questions for me, or guide me to information about this already written. By the way, are there any good forums for discussing Hobbit? I have found a couple forums, but there has been very low activity on them (only a handful of questions, and now replies). /Øyvind

2 3

Monitoring Solaris 10 virtual machines
by annaj＠hi.is 07 Oct '08

07 Oct '08

Solaris 10 has a virtualization capability which is the ability to create so called zones, that effectively act like independent machines. What is the best way to monitor those virtual machines? -- Kindest Regards, Anna Jonna Ármannsdóttir, %& A: Because people read from top to bottom. Unix System Aministration, Computing Services, %& Q: Why is top posting bad? University of Iceland.

4 3

Running scripts in hobbit-alert.cfg
by James.Roberts＠hants.gov.uk 07 Oct '08

07 Oct '08

I am trying to run this from the hobbit-alert.cfg HOST=nim01 SERVICE=disk SCRIPT /home/hobbit/server/ext/emails/emailscript itdcrj at hants.gov.uk COLOR=red REPEAT=1 RECOVERED But it never runs....Any ideas?? James Roberts IT Consultant - IT Services Hampshire County Council Room 415,Elizabeth II Court, The Castle, Winchester, Hants SO23 8US Internal HPSN: (200) 5037 Tel: +44 (0)1962 845037 Fax: +44 (0)1962 847434 james.roberts at hants.gov.uk Web: WWW.Hants.Gov.UK

3 2

BB environment variables
by jon＠shadowsoft.com 07 Oct '08

07 Oct '08

What's a good way to get all the BB environment variables set up when I want to use them in a script that's not launched from hobbitd? For example, if I have a script that receives an email and reports a status because of it? I know that I could hard-code everything, but that's tacky. :-) Thanks, Jon

6 7

no graphs from mpstat
by klomph＠nlr.nl 06 Oct '08

06 Oct '08

I'm trying to setup mpstat to monitor systems with multiple cpu's, but the graphs are not created. Can someone help me with this ? The information is send from the client to the server and in the rrd data directory I have a file called mpstat.CPU2.rrd but there are no values in the last ds (see example) <ds> <name> CPU </name> <type> GAUGE </type> <minimal_heartbeat> 600 </minimal_heartbeat> <min> 0.0000000000e+00 </min> <max> 3.2000000000e+01 </max>  <last_ds> </last_ds> <value> NaN </value> <unknown_sec> 241 </unknown_sec> </ds> the contents of the hobbit_mpstat.hostname.send file is CPUs: 2 CPU0usr: 17.18 CPU0sys: 1.34 CPU0wt: 0.06 CPU0idl: 81.31 CPU1usr: 26.16 CPU1sys: 1.40 CPU1wt: 0.06 CPU1idl: 72.29 And I have copied the rrd_data.pl from the shire and can't find a typo in it. Any suggestions would be appriciated. Regards, Bert Klomp

4 6