October 2008 - Xymon - Xymon mailman web

Server name hyperlinks and oracle
by khemdev＠bc.edu 31 Oct '08

31 Oct '08

Greetings all, I was wondering if there is a way in Hobbit to make the server name hyperlinks (once clicked) connect to oracle database, run a query using the servername as the key and return the results back to the hobbit webpage? I know the bbgen can be passed -docurl=/bb-csvinfo.sh/ to grab info from a csv file, I was wondering if the same thing can be used to run a query on a database? Thanks in advance for the help, Dev Khemraj

2 2

Alerting: Stop After
by annaj＠hi.is 31 Oct '08

31 Oct '08

Hi all, on the info web page for a host, is a column named Stop After. I would like to be able to configure this column. I have not noticed any information about this. See below: Alerting: Service Recipient 1st Delay Stop after Repeat Time of Day Colors conn annaj at hi.is (R) 15m - 4h W:0700:2259 yellow -- 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.

1 0

log monitoring for dmz hosts
by Craig.Cook＠gpi.com 31 Oct '08

31 Oct '08

Never mind. Looks like msgcache does allow updates to the local client. Craig

1 0

bbwin max Message size?
by stewartl42＠gmail.com 31 Oct '08

31 Oct '08

I'm seeing an error with bbwin where the message description does not show up in hobbit. The description alone is about 1600 characters. It looks like this in hobbit... system: warning - 2008/10/30 11:03:13 - NETLOGON (5807) - n/a "" Is there a limit to the description size on individual log entries? -- Stewart -- Why are you unhappy? Because 99.9% Of everything you think, And of everything you do, Is for yourself - And there isn't one.

2 3

log monitoring for dmz hosts
by Craig.Cook＠gpi.com 31 Oct '08

31 Oct '08

Is there a way to allow dmz clients (using msgcache) to monitor non default logs? Normally you specify the log file to monitor in client-local.cfg and what to alert on in client-local.cfg. Next time the client sends data to the server it gets notified about a new log file to start monitoring. Using msgcache this never happens... I can probably do a dodgy trick and define what I want in logfetch.<clientname>.log on the dmz client in its $client/tmp/ dir. Is there a better way? If I change the dmz client to be a "local" client I then have to define thresholds on the local client, I would like to avoid that... Craig

1 0

rrdtool
by jon＠shadowsoft.com 31 Oct '08

31 Oct '08

This isn't quite the place for this, but you guys always give really good answers. :-) If I want to install rrdtool and snmpget on a really "stock" CentOS system, and yum can't find these things in the "usual" places... what do I need to do to get them? Thanks, Jon

3 2

Purple Colours
by Jarrod.Hodder＠paradigmit.net.au 31 Oct '08

31 Oct '08

Hey, I have noticed that on my monitor that a lot of the tests are turning purple. I read in this thread (http://www.hobbitmon.com/hobbiton/2007/08/msg00051.html) that I should check the status of my hobbitclient.sh. I ran the status and it returned. /runclient.sh status Hobbit client (clientlaunch) does not appear to be running So I thought that would be my problem, I navigated to my hobbitclient.sh and tried to run it. This is the error I am getting. root at Hobbits:/usr/lib/hobbit/client/bin# ./hobbitclient.sh ./hobbitclient.sh: 45: /bin/clientupdate: not found ./hobbitclient.sh: 52: /bin/hobbitclient-linux.sh: not found ./hobbitclient.sh: 63: /bin/logfetch: not found ./hobbitclient.sh: 79: @: not found root at Hobbits:/usr/lib/hobbit/client/bin# This is a list of files in my /client/bin/ directory bb bbhostgrep hobbitclient-aix.sh hobbitclient-hp-ux.sh hobbitclient-netbsd.sh hobbitclient-sco_sv.sh hobbitlaunch orcahobbit bbcmd bbhostshow hobbitclient-darwin.sh hobbitclient-irix.sh hobbitclient-openbsd.sh hobbitclient.sh logfetch bbdigest clientupdate hobbitclient-freebsd.sh hobbitclient-linux.sh hobbitclient-osf1.sh hobbitclient-sunos.sh msgcache So from that I can see that the files are infact there, but it can't see them. Any help on where I went wrong would be appreciated. -Jarrod

4 6

wildcards or regex with SPLITNCV
by hobbit＠elyograg.org 30 Oct '08

30 Oct '08

I am working on setting up graphs to track an application. I have created a test that produces the output at the end of this message, I want to use SPLITNCV to handle it, but I don't want to be required to update the hobbit configuration and restart when the number of indexes in the system increases. Is there a way to define the SPLITNCV_testname variable in such a way that it can handle this? Also, I want to be able to set up the live_count and broker_count variables as counters, but be able to track negative changes as well as positive. I'm not sure if the DERIVE format mentioned in the docs will do negative changes. Normally, the value marches ever higher, but when there are problems, it will go down, and I want to see that on the graph. The _time variables and the rest of the _count variables will be gauges. Is this something that can be accomplished? If I have to set up all the graphs as gauge and deal with an ever-increasing value for the special entries, then I will do that. This is on the production system, running version 4.2. I started with the source package from the debian backports archive, then patched with the central BBWin patch and the allinone patch in addition to the existing debian-specific patches. I had to resolve a couple of conflicts, but got it to compile and run. Script output: ============== live_count : 30635889 diff_count : 0 broker_time : 1355 broker_count : 30635889 rss_time : 4525 rss_count : 1680777 indy1s2_time : 863 indy1s2_count : 53532 indy1s3_time : 34 indy1s3_count : 789 indy2s1_time : 23 indy2s1_count : 674006 indy2s2_time : 25 indy2s2_count : 952450 indy3s1_time : 21 indy3s1_count : 1004714 indy4s1_time : 32 indy4s1_count : 997684 indy4s2_time : 32 indy4s2_count : 996035 indy5s1_time : 27 indy5s1_count : 1000602 indy5s2_time : 77 indy5s2_count : 999003 indy6s1_time : 26 indy6s1_count : 1003493 indy6s2_time : 25 indy6s2_count : 1001920 indy7s1_time : 28 indy7s1_count : 999484 indy7s2_time : 26 indy7s2_count : 1003172 indy8s1_time : 25 indy8s1_count : 1004038 indy8s2_time : 24 indy8s2_count : 1000058 indy9s1_time : 25 indy9s1_count : 999745 indy9s2_time : 24 indy9s2_count : 998908 indy10s1_time : 28 indy10s1_count : 995340 indy10s2_time : 25 indy10s2_count : 1001914 indy11s1_time : 21 indy11s1_count : 998153 indy11s2_time : 20 indy11s2_count : 1000448 indy12s1_time : 21 indy12s1_count : 995504 indy12s2_time : 22 indy12s2_count : 996100 indy13s1_time : 26 indy13s1_count : 997570 indy13s2_time : 50 indy13s2_count : 999154 indy14s1_time : 22 indy14s1_count : 997059 indy14s2_time : 21 indy14s2_count : 995457 indy15s1_time : 27 indy15s1_count : 1003780 indy15s2_time : 25 indy15s2_count : 997950 indy16s1_time : 66 indy16s1_count : 999673 indy16s2_time : 34 indy16s2_count : 994129 indy17s1_time : 28 indy17s1_count : 998302 indy17s2_time : 56 indy17s2_count : 975723

3 4

Graph definitions
by Vernon.Everett＠woodside.com.au 30 Oct '08

30 Oct '08

Hi all Is there any way to define a graph definition for values that may or may not be available. Let's work with CPU temperature, just as an example. In hobbitserver.cfg, I can define this NCV_systemp="*:GAUGE" (And the other ncv definitions) It then generates rrd files for all the servers. However, some systems have 1 cpu, some have 2, and some 4. So the rrd files only have entries for the number of CPUs. In the hobbitgraph.cfg we can define the graph, with 4 cpu temperatures, but if the rrd file doesn't contain 4 fields (i.e. system has only 2 CPUs), it doesn't generate a graph. It is also not possible to have a separate graph definition for 1, 2 and 4 CPU systems. The other way round it, is to have each CPU in a separate file, and define FNPATTERN= with a wildcard. But how do we get the ncv module to put each field in a sepreate rrd file? Regards Vernon NOTICE: This email and any attachments are confidential. They may contain legally privileged information or copyright material. You must not read, copy, use or disclose them without authorisation. If you are not an intended recipient, please contact us at once by return email and then delete both messages and all attachments.

3 3

Weird disk errors....the sequel
by URMM＠VM.MARIST.EDU 30 Oct '08

30 Oct '08

I was premature in thinking that Stef's patch solved the problem. However, I've narrowed down the cause. Turns out, it isn't related to the AIX client in general. Its related to one AIX server specifically. This server has a huge number of ports connected, which is a problem we have to deal with, but can't do much about right now. Meanwhile, the number of ports is causing the hobbit server issues. Here's the error messages we get on the hobbitd test: Latest errormessages: Oversize data/client msg from 148.100.11.11 truncated (n=1269789, limit 524288) First line: ldapb|aix|aix| Oversize status msg from 148.100.11.11 for ldapb:ports truncated (n=513615, limit=262144) I think this issue is, in turn, causing the false alerts we are seeing elsewhere. When I remove this server from the bb-hosts config, that problem goes away. Is there a way I can turn off the ports test for this one server? Or, can I limit how much data the client sends to Hobbit server? Any other suggestions? Martha

2 1