centralized management of clientlaunch.cfg 🔗 link

I have a number of custom scripts spread across multiple machines; I
understand how to configure them in clientlaunch.cfg  but I would like
to keep them in a single config file (it's impractical to maintain
this many individual configs). Has anyone written a tool to manage
push/pull of the configs and generating individual hobbit-clients.cfg,
or otherwise duplicate what bb-bbexttab did?

-Lachlan

We maintain everything on one central server and then use rdist to push them
out to the clients. I have one clientlaunch.cfg file which has everything in
it and all of the ext scripts test if they are running on the right host.
This could be done better, but we don't have enough monitored clients to
make it worth the extra effort.

If I was going to improve it I would have the clientlaunch.cfg file have m4
directives to select stanzas for the host being rdisted to (we actually us a
system called distrib built on top of rdist which does this automatically),
that way a single clientlaunch.cfg  file would serve as the source for
multiple files on the various clients. The same system (distrib) would
control which clients got which ext scripts.

Steve.

---

On Wed, Dec 17, 2008 at 11:06 AM, Don Lachlan <user-32e83f32b5ed@xymon.invalid>wrote:

In <user-835fbea2f756@xymon.invalid> "Don Lachlan" <user-32e83f32b5ed@xymon.invalid> writes:

You could grab the "hobbitlaunch" program from the current development
branch; it has a new "ONHOST <pattern>" setting for each task - the
task will only run if the local hostname matches the <pattern>. E.g.

   [myscript]
       CMD $BBHOME/ext/myscript.sh
       ONHOST (www|mail).foo.com


Regards,
Henrik

 
Hi, 
	We're having some trouble getting hobbit to ignore a red proc on
a restart of a process. 

In bb-hosts, I have:

204.155.140.144 iadpimp1-vsnl.interprise.com    #
DOWNTIME=*:0200:0230:systemrestart

In hobbit-clients.cfg, I have:

HOST=iadpimp1-vsnl.interprise.com
        PROC tail 0 1 red "TEXT=tail - testing paging and ignoring
times"
        PROC AppServer.exe 3 7 red "TEXT=Visionael Application Server"

And in hobbit-alerts.cfg, I have:

$HG-PRIMARY=starsky(.*)|iadorap1(.*)|iadnasp1(.*)|mdsobj3(.*)|cher(.*)|h
umility(.*)|stnpres1(.*)|stntrapp1(.*)|mdstf1(.
*)|firetwo(.*)|totlccwm1(.*)|brinkley(.*)|iadcap1(.*)|iadcap2(.*)|stngeo
p1(.*)|iadpres1(.*)|iadpimp1-vsnl(.*)|iadpimp1-
ca(.*)|iadtatp1(.*)|iadobjs1(.*)|phm1ssxc01m
   .
   .
   .
HOST=%$HG-PRIMARY COLOR=red EXSERVICE=memory DURATION>7
        MAIL $NG-IM-ADMIN
        IGNORE HOST=iadpimp1-vsnl(.*) SERVICE=proc
TIME=*:0200:0230,*:1500:1600
        MAIL $NG-IM-ONCALL FORMAT=SMS


We've tried many other combinations. And can get it to work for a test
SERVICE, using bb to make the test service red. 
But when we have procs in there, I get paged no matter what. At 2:15am.
Not so fun.

I put the "PROC tail 0 1 ..." in for testing with proc. 

This happened in 4.2.0. So today, I downloaded 4.2.2 in source forge and
compiled and installed it.  Does the same thing.

I'd also tried the DOWNTIME directive (why you see it there) but that's
not working either. 

Anybody have any words of wisdom here?

Thanks,
Paul.


Paul Root         IM/MNS Infrastructure    


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.

On Wed, Dec 17, 2008 at 4:17 PM, Henrik Størner <user-ce4a2c883f75@xymon.invalid> wrote:

Ooh, I will be on that as soon as I can upgrade.
Thanks!
Steve.