On 10/3/13 9:07 AM, jcbollinger wrote:
And you're certain that $nagios_server evaluates to true on the machines
that nevertheless fail to get the expected Nagios_host resources
applied? You could probe that by changing the above to
>
if $nagios_server {
Nagios_host <<| |>>
notify { 'I am a nagios server': }
} else {
notify { 'I am not a nagios server': }
}
That will also help you catch the case that the class containing the
above is not included in the catalog at all, for then you would see
neither the Notify message in the agent log. You can also watch your
query log for the query by which Puppet retrieves all the wanted
Nagios_host resources; if nothing else is fruitful then the details of
the query might reveal some useful information.
First to answer. I have one computer that runs nagios. That's the
(only) computer that has $nagios_server defined. The computers that I
want to monitor have the @@nagios_host {} resource. I've verified this
up and down.
I've done some more testing, and there's something very funny going on.
In my manifests, I have the following:
notify {"Nagios Base: $fqdn $hostname $ipaddress": }
@@nagios_host { $fqdn:
ensure => present,
alias => $hostname,
address => $ipaddress,
use => "linux-server",
}
On the hosts to be monitored, I see the notify, and, as I've said, I see
the db inserts, but then nothing comes up in the REST query. HOWEVER,
if I take the lines above and put them in a separate .pp file and
"puppet apply" that file, then I see the resources via the REST query.
My understanding was that once a resource gets defined, that's the end
of the story, but it's clearly not what's happening here.
So, "puppet agent --test --environment production" fails to create
resources findable via REST (and in fact removes them if they're
present), but "puppet apply foo.pp" (with foo.pp containing the
statements above) makes them findable (and they show up in
nagios_hosts.cfg on the nagios server).
though I don't expect using that form will change anything. I'm more
interested, though, in the back end. That is, is puppetdb for some
reason failing to serve up resources that are in fact recorded in the
back-end DB, or are the resources either not making it to the DB in the
first place, or are they being purged from the DB between being entered
and being queried?
That's a very good question. I can certainly find catalog_resource and
resource_params records that *look* like they contain appropriate
Nagios_host resources, but the REST API says no, and if I run the actual
sql query out of the postgres statement log, it says no. The sql query
itself is a complex enough to be beyond my db-mojo to understand why
it's not finding any Nagios_host resources.
You said earlier that some nagios hosts are getting configured, at least
on some machines, but I didn't completely understand which hosts are
getting configured on which machines. Are there machines that get all
the expected hosts? Are there nagios_hosts that get configured on all
machines? More generally, what is the nature of the correlation (that
you have recognized) between declared nagios host resources and the
machines on which puppet manages those resources?
Yes, sometimes I've gotten some partial Nagios_host data generated on
the nagios server, but never in any discernible or repeatable form.
It looks like enlightenment lies in understanding the difference between
the puppet agent and puppet apply runs.
--
David Thompson
Waisman Center Brain Imaging and Behavior Lab
1500 Highland Ave. Room T133
Madison, WI 53705-2280
(608) 265-6608
dthompson (at) waisman (dot) wisc (dot) edu
--
You received this message because you are subscribed to the Google Groups "Puppet
Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to puppet-users+unsubscr...@googlegroups.com.
To post to this group, send email to puppet-users@googlegroups.com.
Visit this group at http://groups.google.com/group/puppet-users.
For more options, visit https://groups.google.com/groups/opt_out.
