Thanks for your answer Evan.
For these applications, we need to have a very good control. So we'll
never update them via local update utility nor use an ensure => latest
We always use ensure => x.y.z
We also need to be able to do rollback, so it's simpler to remove
everything from say, version 1
dependency question:
I want to require that Package{ruby-shadow} is installed if its not there on
Redhat. This package is in Yum-repo which has some dependency requires for
files owned by root, so I have created the circular reference
user{root:password} -> Package[ruby-shadow] -> Class["yum"] (h
On Fri, 2009-01-16 at 09:41 -0500, Jason Kohles wrote:
> "MySQL server has gone away" means you lost your connection to mysql,
> most likely because there is a configured idle timeout or the server
> was restarted.
Or, more probably the mysql server has crashed, for any reasons
(including cra
Sorry if I missed this in the documentation...
Does this work according to its design?
When I configure this file with a template and CHANGE the file on the
client server, puppet detects a change and overwrites the file:
file { "jboss-config":
path=> "/etc/sysconfig/jboss",
ensure =
On Wed, Jan 28, 2009 at 3:00 AM, nicolas wrote:
>
>
> Thanks for your answer Evan.
>
> For these applications, we need to have a very good control. So we'll
> never update them via local update utility nor use an ensure => latest
> We always use ensure => x.y.z
> We also need to be able to do rol
Ok, I agree that my request is targeted to a special case that can be
dangerous if applied to general case.
My main concern was that our admins may find tedious to do the two-
steps process for each release of an application. Maybe I should tell
them it's more secure !
Anyway, thanks for sharing
I'm not entirely certain of your scenario, but unless you are going to
purposely have two versions of the same application running in production
for extended periods, this seems like the wrong way to do this.
Do you have some way to roll through Dev, Test, Staging, Production?
Can you just have o
On Wed, Jan 28, 2009 at 6:11 PM, James Turnbull wrote:
>
> -BEGIN PGP SIGNED MESSAGE-
> Hash: SHA1
>
> So I just looked quickly at the doco for pkgsrc and the existing
> freebsd, openbsd and ports support in Puppet and it looks to me like
> it'll work fine. It's currently confined to onl
On Wed, Jan 28, 2009 at 7:26 AM, Luke Kanies wrote:
> Heh, sorry; I was trying to dredge up an appropriate response en
> Français and failed,
Would that be "de rien, et bienvenue"?
--~--~-~--~~~---~--~~
You received this message because you are subscribed to the
Jeff,
What version of Puppet are you using?
I could not reproduce this behavior with either content or source.
When replace was false the file was not changed.
When I changed the file on disk, the logs say that the checksum changed and
that is replaced in the file bucket.
If you can consistent
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Rob Chanter wrote:
>
>
> Can puppet easily enough refer a package resource to an alternative
provider?
>
package { blah:
ensure => latest,
provider => ports
}
The provider attribute allows you to specify which provider to use for
certain functions.
Hi Matt!
On Wed, Jan 28, 2009 at 11:22 AM, Matt McLeod wrote:
> network_interface=primary {hostname=<%= realhostname %>
>
> err: Could not retrieve catalog: Failed to parse template sysidcfg: Could not
> find value for 'realhostname' at
> /staging/puppet/production/manifests/nodes.pp:90 on nod
Hello there
I've written a script that generates a skeleton module, i was going to
drop it onto the recipe page, alas i can't edit that...
#!/bin/bash
moduleDir=`puppetd --configprint modulepath | cut -d ':' -f 1`
module="$1"
moduleInit="$moduleDir/$module/manifests/init.pp"
mkdir -p $moduleD
On Tue, 2009-01-27 at 16:21 +0100, Dan Bode wrote:
> Has anyone used Augeas on solaris? I guess that the lenses may not
> work there.
>
> Has anyone compiled augeas on solaris? I tried, but I got warnings
> aout asprintf not being available and I could not find the
> installation package for it.
Matt,
There is no realhostname variable in scope when the template is getting
called.
Variables in templates either need to be facts or set in the scope of Puppet
they are being called from.
Does that make sense?
Andrew
On Tue, Jan 27, 2009 at 5:22 PM, Matt McLeod wrote:
>
> I'm following t
On Sun, 2009-01-25 at 15:21 -0800, Robin Lee Powell wrote:
> On Mon, Jan 19, 2009 at 01:28:53PM -0800, David Lutterkort wrote:
> >
> >
> > This is what I am trying to address with better path expressions in
> > Augeas itself. That lets you write
> >
> > set "/files/etc/aliases/*[name =
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
We're pleased to announce that Facter release 1.5.3 is now available.
This is largely a maintenance release but there are some new facts and
features. This will be the last release in the 1.5.x branch and the
next release will be 2.0.0.
New facts
OK. So the docco is completely wrong: everything referenced in the
template needs to be set as a variable prior to defining the zone.
I'd assumed from context that the template was going to have the
previously-declared object properties passed to it.
I can live with that, I was going to abstrac
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Matt McLeod wrote:
> OK. So the docco is completely wrong: everything referenced in the
> template needs to be set as a variable prior to defining the zone.
> I'd assumed from context that the template was going to have the
> previously-declared ob
James Turnbull wrote:
>
> -BEGIN PGP SIGNED MESSAGE-
> Hash: SHA1
>
> Matt McLeod wrote:
> > OK. So the docco is completely wrong: everything referenced in the
> > template needs to be set as a variable prior to defining the zone.
> > I'd assumed from context that the template was goi
20 matches
Mail list logo
