Good day,
Thanks for your reply.
> i would have to say your configuration has a lot to do with
> this. if you start
> your services fresh without saving your data you would see
> the initial times
> change, I would suspect.
If we have no data for a host, than the initial times change. That's not a
big deal for me.
> Using the standard config files that come with the nagios
> distribution, the save
> initial state is on,
> this means that when viewing your services you might notice
> that the times are
> way off on new services.
Unless I am way off base, I believe that you completely misread my problem.
Just to make sure, the closest that I could find to "save initial state" is
"log_initial_states", which just logs the initial state when Nagios starts
and is set to "off" in both my example config and the documentation. Are
you referring to something else? If so, what is the actual configuration
option?
> so with the above information, and based on the assumption that your
> configuration is as above, with initial states on, no This is
> not a bug, the
> program is doing what you told it to do.
The setting that most closely matches "save initial states", " is
"log_initial_states". So no, my configuration file doesn't match your
description. But again, I belive that my problem was misunderstood.
Perhaps I can explain the problem again, using a different hypothetical
example.
I start up a web server on a server I'm monitoring. I would like to monitor
the web server. I set up a service to monitor my web server. I configure
it to check every five minutes. I restart Nagios to have it read the new
config. I look on the status .cgi, and see the rest of my services in the
states they were in before I restarted Nagios (which is what I expect). I
see that my new web service is in "Pending" status, waiting for the first
poll to happen (which is what I expect). However, Nagios has scheduled the
first service check for my web service 30 minutes in the future, even though
I asked for it to be checked every 5 minutes (which is NOT what I expect).
I patiently wait for the whole half hour for Nagios to do the first service
check. It gets checked, and then further checks are scheduled every 5
minutes (which is what I expect).
For troubleshooting, I tried to delete the status information and restart
Nagios. They all come up "Pending" (which is what I expect) and Nagios
reschedules all of the checks (which is also what I expect) but it still
does not schedule my 5-minute checks within 5 minutes (which is not what I
expect).
I really don't care that the service isn't polled the instant Nagios starts,
but I certainly would expect the first poll to be done within the time
interval that the service is supposed to be polled in.
============================
Darren Gamble
Planner, Regional Services
Shaw Cablesystems GP
630 - 3rd Avenue SW
Calgary, Alberta, Canada
T2P 4L4
(403) 781-4948
> > > Good day,
> > >
> > > I've discovered what certainly appears to be a bug.
> > >
> > > We have our Nagios machine set to pool almost 200 hosts with
> > > about 1500
> > > services. The machine would not be able to poll everything
> > > every 5 minutes,
> > > so we have our "important" services polled every 5 minutes
> > > and our less
> > > important services (like disk space size) polled less
> > > frequently. It works
> > > well.
> > >
> > > However, we notice a problem when we add brand new services
> > > to hosts- their
> > > first scheduled time is much later than it should be, which I
> > > presume is
> > > related to our unusual configuration. For example, let's say
> > > I have a host
> > > that has some services that are polled every 5 minutes, and
> > > some that are
> > > polled every hour. If I add a new service to the host which
> > > I've scheduled
> > > to be polled every 5 minutes, it may have its first poll time
> > > scheduled in
> > > 32 minutes (i.e. it misses its polls 2, 7, 12, 17, 22 and 27
> > > minutes from
> > > the program start).
> > >
> > > After the service is finally polled for the first time,
> it
...[email truncated]...
This post was automatically imported from historical nagios-devel mailing list archives
Original poster: Darren.Gamble@sjrb.ca