[hobbit] hobbit-clients configuration not taking

[HUBBARD, GREG ((EDS))]

re: client update -- it is a murky subject.  It is documented somewhere, but I forget where.  what you do is tar up your client directory and compress it, and then put it somewhere. then you can refer to it in the client-local.cfg file.  When the client checks in it will check to see if there is a new version.  If so, it downloads it and installs it (by unpacking the tar ball).  This works for Solaris, but does not work for AIX.  I would use it carefully.  This also means that you have to deal with each host individually in your client-local file, but this is not a bad thing unless you have thousands of them.

GLH

________________________________
From: Josh Luthman [mailto:josh at imaginenetworksllc.com]
Sent: Wednesday, September 09, 2009 3:24 PM
To: hobbit at hswn.dk
Subject: Re: [hobbit] hobbit-clients configuration not taking

Confirmed - PROCS works on the host in which the server is getting client data from.

I compiled on each client individually (see the last link of my last post).  Exact same commands on each - literally copy/paste short of root passwords.  I will recompile the two failing clients tonight to see what happens.

I wasn't aware of any client update.  Assuming this is not default, but still, is there a man page detailing this?

Josh Luthman
Office: 937-552-2340
Direct: 937-552-2343
1100 Wayne St
Suite 1337
Troy, OH 45373

"When you have eliminated the impossible, that which remains, however improbable, must be the truth."
--- Sir Arthur Conan Doyle


On Wed, Sep 9, 2009 at 4:07 PM, Greg Hubbard <glh.forums at gmail.com<mailto:glh.forums at gmail.com>> wrote:
Good.  So if you tried PROC on "two" you should get something.  Something is deeply wrong on the other two hosts.  And Xymon likes to silently ignore errors.

My **guess** is that something crummy has happened to the failing clients.  Since there is not much to them they should be easy enough to check out.  Might be a good education.

Did you install from RPM, or compile your own copy of the client?  Are you able to reinstall the client piece from a known working source?  Are you using the "client update" piece where the client can fetch a tarball from the server and install it on its own?

GLH


On 9/9/09, Josh Luthman <josh at imaginenetworksllc.com<mailto:josh at imaginenetworksllc.com>> wrote:
Suggesting that the bbc is capable of labeling tests.  The hobbit-clients' data is staying clear.

two.myhost.com<http://two.myhost.com/> does have data and it is [ps] and [msgs:/var/log/messages]

one.myhost.com<http://one.myhost.com/> and three.myhost.com<http://three.myhost.com/> has no data on the server.

hobbitclient.log is full of:
2009-09-09 14:38:44 Failed to get a message, terminating

$bb server.fqdn.com<http://server.fqdn.com/> ping #this command works
hobbit-4.2.0

This is running a CentOS release 5.2 (Final) and CentOS release 5 (Final) respectively - "Red Hat Linux".

One thread on the archives was able to fix it by installing the Debian package.  Obviously I can't do that.  The strange part is I executed the exact same commands across all three - the same ones I put up here
http://en.wikibooks.org/wiki/System_Monitoring_with_Xymon/Administration_Guide/Compiling_on_CentOS

Perhaps there is a package missing of some sort?  Is there a way to get more detailed logs from the hobbitclient.log?

Josh Luthman
Office: 937-552-2340
Direct: 937-552-2343
1100 Wayne St
Suite 1337
Troy, OH 45373

"When you have eliminated the impossible, that which remains, however improbable, must be the truth."
--- Sir Arthur Conan Doyle


On Wed, Sep 9, 2009 at 3:34 PM, Greg Hubbard <glh.forums at gmail.com<mailto:glh.forums at gmail.com>> wrote:
Then what was this all about?  I guess I read it wrong.

"However on an bbc 1.9i btf (better then free) client it is pushing PROCs to the server:

#cat ~bbuser/bbc1.9i-btf/etc/bbwarnsetup.cfg:
...
PROCS="bbrun"                           # (YELLOW) WARN IF NOT RUNNING
PAGEPROC="cron"                         # (RED) PAGE IF NOT RUNNING
export PROCS PAGEPROC"
What you might want to do is track down the "raw data" being sent by these hosts to see if they contain a PROC secion.  I think you can get to it on each host through the http://<xymon-server>/hobbit-cgi/bb-hostsvc.sh?CLIENT=<client<http://%3cxymon-server%3e/hobbit-cgi/bb-hostsvc.sh?CLIENT=%3Cclient>> link.  On Solaris you would look for a [ps] header, then a list of processes.  This will show you that the data is getting pumped over.  (For Linux, etc. there might be something different).

And, if your config fragment is correct, you would not expect any PROC checks for one or two, just for three, since you only defined them for three and there are no PROC checks in the DEFAULT section.




--
Disclaimer:  1) all opinions are my own, 2) I may be completely wrong, 3) my advice is worth at least as much as what you are paying for it, or your money cheerfully refunded.

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.xymon.com/pipermail/xymon/attachments/20090909/5a2bab58/attachment.html>