Puppet users - Feb 2009 - Puppet Idioms for distributing RPMs and/or buildign source tarballs?

Sorry for the topic confluence, but I was wondering if there were
common puppet "idioms" for either:

  - distributing an RPM to a series of clients (namely RabbitMQ) and
installing the package?  I realize that I can do it naively by pulling
the file from the puppetmaster and then using the package resource to
install it via RPM, but I was wondering if there was a better way to
do this.  I would love to be able to install via RPM with a URL (i.e.
rpm -Uvh http://....), but the command that actually gets executed
with puppet (something like ''rpm -i --oldpackage http://....)
doesn''t
seem to be idempotent, i.e. it fails after the package is already
installed.  Also, I would rather not have all of my puppets calling
out to the rpm url every 2 minutes.

  - distributing and compiling a tarball?  Its a pretty standard
autoconf/automake tarball (i.e. ''./configure --prefix=/usr; make; make
install'') ... is there a "standard" way to do this with
puppet.  I''m
also using a source-built Ruby and RubyGems, which would be nice to
manage with puppet.

I''ve looked through a bunch of existing modules, but I haven''t
seen
this sort of thing.

Thanks for any help,
-Mike

--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups
"Puppet Users" group.
To post to this group, send email to puppet-users@googlegroups.com
To unsubscribe from this group, send email to
puppet-users+unsubscribe@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/puppet-users?hl=en
-~----------~----~----~----~------~----~------~--~---

In response to your first comment, it certainly should be idempotent.
Puppet should be able to determine that the rpm is already installed,
and I believe that is just a local "rpm -q" or something. So, I
don''t
think it should hit the URL unless it doesn''t think it''s
installed.
This can happen, though, when the name of the resource doesn''t match
the name that "rpm -qa" returns. Perhaps you can provide the output of
one of the non-idempotent puppet runs with the --debug and --trace
options.

I also occasionally use something like this for packages that have a
different name when you install vs when you query. FreeBSD is full of
this:
  http://pastie.org/397847

What''s not clear to me is how to tell rpm to do an "-U"
update instead
of an "-i" install. I believe I have used execs to get around this in
the past.

As far as building from source, I just create a puppet definition like
so:
  http://pastie.org/397821

And I use it like this:
  http://pastie.org/397827

If there is a better way out there, I''m all ears! :)

Anyways, I hope that helped in some way.

 - Jeremy


On Feb 23, 7:40 am, "Michael L. Artz" <mla...@gmail.com>
wrote:
> Sorry for the topic confluence, but I was wondering if there were
> common puppet "idioms" for either:
>
>   - distributing an RPM to a series of clients (namely RabbitMQ) and
> installing the package?  I realize that I can do it naively by pulling
> the file from the puppetmaster and then using the package resource to
> install it via RPM, but I was wondering if there was a better way to
> do this.  I would love to be able to install via RPM with a URL (i.e.
> rpm -Uvh http://....), but the command that actually gets executed
> with puppet (something like ''rpm -i --oldpackage http://....)
doesn''t
> seem to be idempotent, i.e. it fails after the package is already
> installed.  Also, I would rather not have all of my puppets calling
> out to the rpm url every 2 minutes.
>
>   - distributing and compiling a tarball?  Its a pretty standard
> autoconf/automake tarball (i.e. ''./configure --prefix=/usr; make;
make
> install'') ... is there a "standard" way to do this with
puppet.  I''m
> also using a source-built Ruby and RubyGems, which would be nice to
> manage with puppet.
>
> I''ve looked through a bunch of existing modules, but I
haven''t seen
> this sort of thing.
>
> Thanks for any help,
> -Mike
--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups
"Puppet Users" group.
To post to this group, send email to puppet-users@googlegroups.com
To unsubscribe from this group, send email to
puppet-users+unsubscribe@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/puppet-users?hl=en
-~----------~----~----~----~------~----~------~--~---

On 2/23/2009 9:40 AM, Michael L. Artz wrote:
>   - distributing an RPM to a series of clients (namely RabbitMQ) and
> installing the package?  I realize that I can do it naively by pulling
> the file from the puppetmaster and then using the package resource to
> install it via RPM, but I was wondering if there was a better way to
> do this.
Make a regular http- or ftp-accessible repository (instead of using the 
puppetmaster''s file distribution facility), and use one of the 
higher-level rpm-based package providers with the puppet package type 
(yum, aptrpm, etc.). You might also be able to do this just with rpm and 
a repository (I''m Redhat-illiterate). But the important thing would be 
to ensure you can check if a package is already installed before you 
attempt another rpm installation. Even if you had to fall back to an

   exec { "install-rabbit":
     command => "rpm -Uvh http://path/to/rabbit.rpm"
   }

you could make that idempotent with an unless parameter:

   exec { "install-rabbit":
     command => "rpm -Uvh http://path/to/rabbit.rpm",
     unless  => "rpm -qsomethingsomething rabbit | grep -q
installed"
   }
>   - distributing and compiling a tarball?  Its a pretty standard
> autoconf/automake tarball (i.e. ''./configure --prefix=/usr; make;
make
> install'') ... is there a "standard" way to do this with
puppet.  I''m
> also using a source-built Ruby and RubyGems, which would be nice to
> manage with puppet.
I use GNU stow and a stowedpackage definition: 
http://blogs.cae.tntech.edu/mwr/2008/02/01/the-autostow-is-dead-long-live-stowedpackage/
-- this wouldn''t exactly work for things you insist on building per 
node, but could be modified. It''d be very similar to what Jeremy
pointed
to in the other reply.

-- 
Mike Renfro  / R&D Engineer, Center for Manufacturing Research,
931 372-3601 / Tennessee Technological University

--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups
"Puppet Users" group.
To post to this group, send email to puppet-users@googlegroups.com
To unsubscribe from this group, send email to
puppet-users+unsubscribe@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/puppet-users?hl=en
-~----------~----~----~----~------~----~------~--~---

Hi
> Make a regular http- or ftp-accessible repository (instead of using the 
> puppetmaster''s file distribution facility), and use one of the 
> higher-level rpm-based package providers with the puppet package type 
> (yum, aptrpm, etc.). You might also be able to do this just with rpm and 
> a repository (I''m Redhat-illiterate). But the important thing
would be
> to ensure you can check if a package is already installed before you 
> attempt another rpm installation. Even if you had to fall back to an
> 
>    exec { "install-rabbit":
>      command => "rpm -Uvh http://path/to/rabbit.rpm"
>    }
> 
> you could make that idempotent with an unless parameter:
> 
>    exec { "install-rabbit":
>      command => "rpm -Uvh http://path/to/rabbit.rpm",
>      unless  => "rpm -qsomethingsomething rabbit | grep -q
installed"
>    }
package{''rabbit'':
    ensure => present,
    source => "http://path/to/rabbit.rpm",
    privater => rpm,
}

works as well.
However managing your own yum repository is easy and imho the best way
to go.

cheers pete

--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups
"Puppet Users" group.
To post to this group, send email to puppet-users@googlegroups.com
To unsubscribe from this group, send email to
puppet-users+unsubscribe@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/puppet-users?hl=en
-~----------~----~----~----~------~----~------~--~---

Sourcing to a HTTP address works? Since when?

On Mon, Feb 23, 2009 at 4:45 PM, Peter Meier <peter.meier@immerda.ch>
wrote:
>
> Hi
>
> > Make a regular http- or ftp-accessible repository (instead of using
the
> > puppetmaster''s file distribution facility), and use one of
the
> > higher-level rpm-based package providers with the puppet package type
> > (yum, aptrpm, etc.). You might also be able to do this just with rpm
and
> > a repository (I''m Redhat-illiterate). But the important thing
would be
> > to ensure you can check if a package is already installed before you
> > attempt another rpm installation. Even if you had to fall back to an
> >
> >    exec { "install-rabbit":
> >      command => "rpm -Uvh http://path/to/rabbit.rpm"
> >    }
> >
> > you could make that idempotent with an unless parameter:
> >
> >    exec { "install-rabbit":
> >      command => "rpm -Uvh http://path/to/rabbit.rpm",
> >      unless  => "rpm -qsomethingsomething rabbit | grep -q
installed"
> >    }
>
> package{''rabbit'':
>    ensure => present,
>    source => "http://path/to/rabbit.rpm",
>    privater => rpm,
> }
>
> works as well.
> However managing your own yum repository is easy and imho the best way
> to go.
>
> cheers pete
>
> >
>
--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups
"Puppet Users" group.
To post to this group, send email to puppet-users@googlegroups.com
To unsubscribe from this group, send email to
puppet-users+unsubscribe@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/puppet-users?hl=en
-~----------~----~----~----~------~----~------~--~---

Hi
> Sourcing to a HTTP address works? Since when?
always.

package is not file.

cheers pete

--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups
"Puppet Users" group.
To post to this group, send email to puppet-users@googlegroups.com
To unsubscribe from this group, send email to
puppet-users+unsubscribe@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/puppet-users?hl=en
-~----------~----~----~----~------~----~------~--~---

Jeremy Pruitt wrote:
> In response to your first comment, it certainly should be idempotent.
> Puppet should be able to determine that the rpm is already installed,
> and I believe that is just a local "rpm -q" or something. So, I
don''t
> think it should hit the URL unless it doesn''t think it''s
installed.
> This can happen, though, when the name of the resource doesn''t
match
> the name that "rpm -qa" returns. Perhaps you can provide the
output of
> one of the non-idempotent puppet runs with the --debug and --trace
> options.
>
> I also occasionally use something like this for packages that have a
> different name when you install vs when you query. FreeBSD is full of
> this:
>   http://pastie.org/397847
>
> What''s not clear to me is how to tell rpm to do an "-U"
update instead
> of an "-i" install. I believe I have used execs to get around
this in
> the past.
>
> As far as building from source, I just create a puppet definition like
> so:
>   http://pastie.org/397821
>
> And I use it like this:
>   http://pastie.org/397827
>
> If there is a better way out there, I''m all ears! :)
>
> Anyways, I hope that helped in some way.
>
>  - Jeremy
>
>
> On Feb 23, 7:40 am, "Michael L. Artz" <mla...@gmail.com>
wrote:
>   
>> Sorry for the topic confluence, but I was wondering if there were
>> common puppet "idioms" for either:
>>
>>   - distributing an RPM to a series of clients (namely RabbitMQ) and
>> installing the package?  I realize that I can do it naively by pulling
>> the file from the puppetmaster and then using the package resource to
>> install it via RPM, but I was wondering if there was a better way to
>> do this.  I would love to be able to install via RPM with a URL (i.e.
>> rpm -Uvh http://....), but the command that actually gets executed
>> with puppet (something like ''rpm -i --oldpackage http://....)
doesn''t
>> seem to be idempotent, i.e. it fails after the package is already
>> installed.  Also, I would rather not have all of my puppets calling
>> out to the rpm url every 2 minutes.
>>
>>   - distributing and compiling a tarball?  Its a pretty standard
>> autoconf/automake tarball (i.e. ''./configure --prefix=/usr;
make; make
>> install'') ... is there a "standard" way to do this
with puppet.  I''m
>> also using a source-built Ruby and RubyGems, which would be nice to
>> manage with puppet.
>>
>> I''ve looked through a bunch of existing modules, but I
haven''t seen
>> this sort of thing.
>>
>> Thanks for any help,
>> -Mike
>>     
> >
>   
I wrote a definition when I first started to use puppet for a 
source_package that puts it under /usr/src then subsequently builds it 
based on a string of cfg_opts and make_opts, and uses onlyif to serve 
out and execute the build based on a test or two... it''s a little hairy
and usually best to distribute things through the distro''s PM, but I 
wanted to do it as an academic exercise more than anything. I can try to 
unearth it if you express interest.

--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups
"Puppet Users" group.
To post to this group, send email to puppet-users@googlegroups.com
To unsubscribe from this group, send email to
puppet-users+unsubscribe@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/puppet-users?hl=en
-~----------~----~----~----~------~----~------~--~---