January 2011 - Xymon - Xymon mailman web

performance alerts
by troy＠athabascau.ca 14 Jan '11

14 Jan '11

I'm trying to figure out how to get "bad performance" alerts. Here is my bb-hosts entry: 192.168.10.121 mbox1 # imaps I want to get an alert when the response time goes above 100 ms. Any thougths? Troy Troy Adams Athabasca University __ This communication is intended for the use of the recipient to whom it is addressed, and may contain confidential, personal, and or privileged information. Please contact us immediately if you are not the intended recipient of this communication, and do not copy, distribute, or take action relying on it. Any communications received in error, or subsequent reply, should be deleted or destroyed. ---

4 5

Problem with file owner/group check
by ian.marsh＠hants.gov.uk 14 Jan '11

14 Jan '11

We recently encountered a problem with directory permissions on one of our web servers and so decided to update our Xymon installation (v4.2.3) to monitor it. Unfortunately I think I've found a bug in the code that facilitates this. In the hobbit-client.cfg file we have a test like this: FILE /var/somefile mode=755 owner=root groupid=root yellow The owner test works fine but the group test doesn't, instead it say that the file should be owned by group 'd=root'. Further investigation shows that the owner test also fails in the same way if we use the 'ownerid=root' variation of the owner test. It's almost as if the code is assuming that the first 6 characters of the token is the name and everything else is the value. This works fine with the 'owner=root' variation but fails dismally with the 'ownerid=root' and 'groupid=root' variations. It would be fine if we could use 'group=root' but as that's a reserved word it's not allowed. Unfortunately my C coding is virtually non-existent, I can get by with reading it but not writing it! Has anyone encountered this before and written a patch for it? Or has it been fixed in the 4.3.0 beta? Thank you, Ian Marsh IT Service Operations - Network Services

2 4

xymon-4.3.0.beta3 on ubuntu 9.10
by lars.ebeling＠leopg9.no-ip.org 13 Jan '11

13 Jan '11

Dear all, during configuration script I get this: Checking for LDAP ... (Open)LDAP include- or library-files not found. If you want to perform detailed LDAP tests (queries), you need to install LDAP an LDAP client library that Xymon can use. OpenLDAP can be found at http://www.openldap.org/ If you have OpenLDAP installed, use the "--ldapinclude DIR" and "--ldaplib DIR" options to configure to specify where they are. Continuing with LDAP support disabled. That's OK, but when compiling I get: gcc -g -O2 -Wall -Wno-unused -D_REENTRANT -D_LARGEFILE_SOURCE -D_FILE_OFFSET_BIT S=64 -DLINUX -I. -I/home/hobbit/xymon-4.3.0.beta3/include -I/usr/include -c -o ldaptest.o ldaptest.c ldaptest.c: In function Ôadd_ldap_testÔ: ldaptest.c:99: error: ÔbadurlÔ undeclared (first use in this function) ldaptest.c:99: error: (Each undeclared identifier is reported only once ldaptest.c:99: error: for each function it appears in.) ldaptest.c:101: error: ÔreqÔ undeclared (first use in this function) ldaptest.c:112: error: ÔbasecheckÔ undeclared (first use in this function) makeÄ1Å: *** Äldaptest.oÅ Error 1 makeÄ1Å: Leaving directory é/home/hobbit/xymon-4.3.0.beta3/xymonnet' make: *** Äxymonnet-buildÅ Error 2 -- Regards Lars Ebeling Rentier http://leopg9.no-ip.org "I am not young enough to know everything." -- Oscar Wilde

4 3

How to create own XYMON custom page like "Views" -> "All non-green view" in www.xymon.com
by jayantak007＠gmail.com 13 Jan '11

13 Jan '11

Hi, Once i open the URL: www.xymon.com and select the menu option "Views" -> "All non-green view", then we can see the following page: 1 events received in the past 126 minutes Wed Jan 5 06:16:03 2011 osiris.hswn.dk conn [image: clear]</xymon-cgi/historylog.sh?HOST=osiris.hswn.dk&SERVICE=conn&TIMEBUF=Tue_Jan_4_23:55:23_2011> [image: >From -> To] [image: green]</xymon-cgi/historylog.sh?HOST=osiris.hswn.dk&SERVICE=conn&TIMEBUF=Wed_Jan_5_06:16:03_2011> I would like to create similar type of display in my own XYMON page. For exameple: Pre-paid Invalid Object (xxx) | SMS (xxx) | Provisioning (xxxx) [image: Invalid_Objects:red:33m]<http://192.168.3.12/xymon-cgi/bb-hostsvc.sh?HOST=HS21-A&SERVICE=Invalid_Obj…> [image: Invalid_Objects:red:33m]<http://192.168.3.12/xymon-cgi/bb-hostsvc.sh?HOST=HS21-A&SERVICE=Invalid_Obj…> [image: Invalid_Objects:red:33m]<http://192.168.3.12/xymon-cgi/bb-hostsvc.sh?HOST=HS21-A&SERVICE=Invalid_Obj…> Where (xxx) is the number of Invalid Object/SMS/Provisioning etc. It is nothing but simple SQL query output. Kindly let me know steps(file name, parameter name etc. need to be configure) to achieve the above. With Regards, Jayanta.

4 6

False alarms due to corrupt client data?
by tkaehn＠netcologne.de 13 Jan '11

13 Jan '11

Hi, during the last month our Xymon server (Debian package from backports 4.3.0~beta2.dfsg-5~bpo50+1.1) reports false alarms from time to time. The alarm page always shows a couple of ] inserted into the clients output: ... Mon Jan 10]15:00:07 CET 2011 - Filesystems NOT ok red 16% / (54448692% used) has reached the PANIC level (90%) red 0% /lib/init/rw (1742112% used) has reached the PANIC level (90%) red 8% /dev (9496% used) has reached the PANIC level (90%) red 0% /dev/shm (1742112% used) has reached the PANIC level (90%) Filesystem ]1024-blocks Used Available Capacity Mounted on /dev/sda2 68025748 10121476 54448692 16% / tmpfs 1742112 0 1742112 0% /lib/init/rw udev 10240 744 9496 8% /dev tmpfs 1742112 0 1742112 0% /dev/shm ... In contrast to this the client data file looks fine: ... [who] [df] Filesystem 1024-blocks Used Available Capacity Mounted on /dev/sda2 68025748 10121476 54448692 16% / tmpfs 1742112 0 1742112 0% /lib/init/rw udev 10240 744 9496 8% /dev tmpfs 1742112 0 1742112 0% /dev/shm [mount] ... I've seen that a couple of Xymon users reported similar errors from time to time. However I didn't find a solution. We already tried to increase the data/message size, but this didn't solve the problem. Best regards Thomas Kähn -- Thomas Kähn Technik, Network Engineering & Design; Content Delivery Platform & IP --------------------------------------------------------------------- NETCOLOGNE Gesellschaft für Telekommunikation mbH Am Coloneum 9 | 50829 Köln Tel: 0221 2222-8718 | Fax: 0221 2222-78718 www.netcologne.de Geschäftsführer: Dipl.-Ing. Karl-Heinz Zankel HRB 25580, AG Köln Diese Nachricht (inklusive aller Anhänge) ist vertraulich. Sollten Sie diese Nachricht versehentlich erhalten haben, bitten wir, den Absender (durch Antwort-E-Mail) hiervon unverzüglich zu informieren und die Nachricht zu löschen. Die E-Mail darf in diesem Fall weder vervielfältigt noch in anderer Weise verwendet werden.

2 2

New installation - differing behavior for 4.2.3 compared to 4.2.2
by Paul.Root＠qwest.com 12 Jan '11

12 Jan '11

Hi all, I've been using xymon/hobbit for a few years now and am building a new system on new hardware, with xymon 4.2.3 on CentOS. I've got a few issues that are messing me up. The first two are things that are different from the old system, which was hobbit and then I upgraded to xymon 4.2.2 on Solaris 10 Sparc). 1. Fully qualified domain names. In my old system, the majority of our machines are identified with just the hostname. There are a few that are fqdn. And on the web pages, those machines display the FQDN. On the new system the domain name is truncated. We don't want that. We have it that way for a reason, and need it to be displayed. FQDN is TRUE on both old and new. I'm not seeing much else in hobbitserver.cfg that would relate. The way it's explained in the documentation, I don't think changing it to FALSE will do what I want, but I may be wrong. 2. The proxy server is not proxying all tests. Things like comm, ntp, and dns tests are not forwarding. We have some machines that are not directly accessible to the primary server, and so the proxy has ntp, dns, http, etc in the comment field of bb-hosts. Those aren't in the main servers bb-hosts file. My new proxy and new main server are reversed from the old ones. That is, the new server is on the main network, where the old was in the private network. The proxies are also reversed. On the old system, the proxy handled conn, dns, ntp, etc just fine and they came up on the server. Again, I need those tests on the main server. 3. The last is related to the proxy server and is new behavior. I'd really like to have my proxy server be a backup for the main server. If the main server goes down, it would see that and take over sending alerts. I do that today with a secondary server to the old hobbit server. It's just a simple script that moves hobbit-alerts.cfg around as needed. Anyway, is there a way to have the proxy server also run hobbitd? Maybe on a different port? I suppose I could bring up another virtual machine but, I don't really want to do that. That's all I got for now. Thanks, Paul. Paul Root Lead Internet Systems Eng Network Services This communication is the property of Qwest and may contain confidential or privileged information. Unauthorized use of this communication is strictly prohibited and may be unlawful. If you have received this communication in error, please immediately notify the sender by reply e-mail and destroy all copies of the communication and any attachments.

2 2

holes in trend graphs
by dominique.frise＠unil.ch 12 Jan '11

12 Jan '11

Hi, This is probably a duplicate of http://www.xymon.com/archive/2010/01/msg00139.html We are running xymon-4.3.0.beta3. The "CPU Load" and "Users and Processes" graphs of the trends column show random 10min holes (see attached holes.png). The other graphs are OK. We see this behaviour on all Solaris and MacOSX clients; Linux and Windows clients are NOT affected. There is nothing in clientlaunch.log (client) or rrd-status.log (server) Dominique

3 3

Installing another xymon instance with MYSQL DB
by shai.feldman＠rsa.com 12 Jan '11

12 Jan '11

Hi I have a running Xymon 4-2-3 on a Linux server . I would like to do the followings and I need help on how to do it . 1.I would like to install additional Instance on the existing server . 2.I would like that the new Instance will store his data on a MYSQL DB instead or in parallel to the RRD DB . The MYSQL server will be installed on the same XYMON server . Please advise . Thanks Shai Feldman | Senior Unix SysAdmin | RSA, The Security Divison of EMC Office +972 737-807-147 | Cell +972 543 343 068 | Fax +972 9972 8101 www.rsa.com<http://www.rsa.com/>

2 1

Setting LANG=C as default?
by hj.beckers＠kreis-steinfurt.de 12 Jan '11

12 Jan '11

At least xymonreports.sh needs LANG set to C or en_SOMETHING. The following commands su - xymon -c " /home/xymon/server/bin/xymoncmd --env=/home/xymon/server/etc/xymonserver.cfg xymonreports.sh monthly" su - xymon -c " LANG=C /home/xymon/server/bin/xymoncmd --env=/home/xymon/server/etc/xymonserver.cfg xymonreports.sh monthly" give the following output: start /home/xymon/server/bin/xymonreports.sh date: ungültiges Datum 12 Dez 2010 00:00:00 2011-01-12 08:33:36 Invalid --reportopts option: Must have start- and end-times Ende /home/xymon/server/bin/xymonreports.sh start /home/xymon/server/bin/xymonreports.sh Ende /home/xymon/server/bin/xymonreports.sh date generates the german 3-letter abbreviation "Dez" for december. Setting LANG to "C" gives the english "dec". Yours hjb

1 0

Here is a hack if you want to use xymon-client to monitor the number of files in a directory
by spa＠syntec.co.uk 11 Jan '11

11 Jan '11

Although there has been discussion in the past (http://www.xymon.com/archive/2007/02/msg00047.html) about a way to track numbers of files in a directory through the xymon client, no working solution was posted (and the only one I found (http://www.xymon.com/archive/2007/02/msg00079.html) the author claimed it would work, but (a) it does not - variable should not have a $ before it when you set it, and (b) the author gives no indication of how it should be used). I had a use for this facility, and didn't need the normal dir count that returns a size in KB (or whatever du returns on your server), and did not want to write an ext script as that seemed to create unnecessary load, so after several false starts, and finally having to read the source code, I found a working solution, which was to write a script that output the information in the same format as the du command and then redefine the du command in xymonclient.cfg as per the two listings below. Create filecount.sh in the client/bin folder, chmod +x it, and replace the du command by setting the DU environment variable in client/etc/xymonclient.cfg... [xymon at myServer client]$ cat bin/filecount.sh #!/bin/sh # Written by SebA 11/01/11 #filecount=`ls $1 | wc -l` # The following is more efficient, at least on CentOS 5: 'it will show files immediately and it has no memory limitations.' #filecount=`ls -1 -U $1 | wc -l` # Must be >=1, or Xymon will interpret the result as a failure, so include . and .. (and 'hidden') files. filecount=`ls -a -1 -U $1 | wc -l` # Ensure we have a tab between the number and the pathname... echo "$filecount $1" That last line may have been converted from a tab character, so replace the consecutive space characters when you use it! [xymon at myServer client]$ grep filecount etc/xymonclient.cfg DU="$XYMONHOME/bin/filecount.sh" # Replace with my filecount script. /Seb You can then use dir as normal on the server in client-local.cfg and hobbit-clients.cfg (sorry - my client is running newest xymon, while the server is running hobbit). So you can use SIZE on the file-count where it would normally work on the dir-size (du). ---------------------------- Of course, proper client/server support, instead of overriding du would be better, but maybe this will help someone until that happens, or maybe this will be the trigger for that to happen. ;) In the course of my debugging, I was somewhat frustrated by the length of time changes took to progagate from the server to the client (but I can live with this). What would be routinely useful though is that even if the result is a green in the files column, that it displayed what the criteria are (like it does if the result is non-green)? Because it (hobbit-clients.cfg) is clearly cached by the server and not checked when the client reports. (I believe hobbit-alerts.cfg (or whatever it is called now) is checked when the client reports, no? So they do not behave the same.) Anyway, this caching means one does not know which criteria were used to determine a green result. Plus, displaying it in the page would be more convenient as most people don't have the file open in another window! Kind regards, SebA Internal Virus Database is out of date. Checked by AVG - www.avg.com Version: 8.5.449 / Virus Database: 271.1.1/3243 - Release Date: 11/07/10 19:34:00

1 0