What Nan said. Personally, I would recommend just renaming nodes.pp to
site.pp -- there's not much of a reason to maintain your nodes in a
separate file when you're just getting started. And yeah, always have
a
node default {
...
}
, even if all it does is fire a notify resource saying "Hey, I'
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
> I've seen excessive memory usage in some clients due to odd variants -
> the Linux OOM killer usually kicks in at some point. Check 'dmesg' to
> make sure it wasn't this.
>
> Also - are you catching core dumps on your box incidentally? And as
> Tore
I've seen excessive memory usage in some clients due to odd variants -
the Linux OOM killer usually kicks in at some point. Check 'dmesg' to
make sure it wasn't this.
Also - are you catching core dumps on your box incidentally? And as
Tore mentions - anything in the logs?
Of course the obvious th
Does your logs say anything about puppetd dieing during the network
outage?
On 7 apr, 07:09, Nat wrote:
> Hi list,
>
> has anyone else experienced puppet clients that die when the network goes
> down for an extended period of time ?
>
> OS rhel 5
> puppet clients / master 0.25.1
>
> clients are l
Yes I can. The server will have a LOT of log info as it's serving up
just under 700 clients, but I can pick a few clients and enable debug
logging for them and report back.
On Oct 2, 2:32 pm, "Andrew Shafer" <[EMAIL PROTECTED]> wrote:
> Can you get debug logs on the client and server?
>
> On Thu
Can you get debug logs on the client and server?
On Thu, Oct 2, 2008 at 8:25 AM, josh <[EMAIL PROTECTED]> wrote:
>
> Paul,
>
> Yes puppetmasterd has been restarted several times but no luck. I'm
> upgrading all the clients to facter-1.5.2 now, and once that done I
> may either upgrade all the cl
Paul,
Yes puppetmasterd has been restarted several times but no luck. I'm
upgrading all the clients to facter-1.5.2 now, and once that done I
may either upgrade all the clients to puppet-0.24.5, or downgrade the
server to 0.24.4, that way both the clients and the server will have
the same versio
Have you restarted puppetmasterd? Often when I see changes not
propagating to clients, it turns out there was a syntax error which
stopped the puppetmaster from reloading changes.
Try restarting puppetmasterd and watch the logs.
--Paul
On Mon, Sep 29, 2008 at 7:29 PM, josh <[EMAIL PROTECTED]> w
I believe the problem is on the client. Debugging logs there would be the
place I'd start.
On Mon, Sep 29, 2008 at 8:29 PM, josh <[EMAIL PROTECTED]> wrote:
>
> The clients still check in every 30 minutes or so, but aren't
> downloading the new classes from the server, and they wil still be
> tryi
The clients still check in every 30 minutes or so, but aren't
downloading the new classes from the server, and they wil still be
trying to download non-existant classes. I'm pushing out an upgrade
of facter from 1.3.8 to 1.5.2, as of right now 500 out of 700 hosts
have the new facter version, the
What do the logs look like on the clients that stop connecting?
That's where I'd expect to see something, not on the master.
On Mon, Sep 29, 2008 at 11:14 AM, josh <[EMAIL PROTECTED]> wrote:
>
> Here's the scenario,
>
> We have roughly 700 OpenSolaris hosts running puppet-0.24.4,
> facter-1.3.8
11 matches
Mail list logo
