James Turnbull schrieb:
> David Schmitt wrote:
>> Meanwhile I'm also trying to merge the various forks and advancements of
>> the modules. The first batch of modules can be seen on my github
>> account[3].
>
> I look forward to all the people who manage module collections putting
> in their ten
What's the correct yum repo to use for installing Puppet & Facter on
RHEL5 and Centos5?
I used to get them from the dlutter-rhel5 repo but this seems to be
massively out of date now - latest version of puppet-server in there is
0.24.5-1.el5 and facter 1.5.4-1.el5.
In Epel I see puppet-server 0.2
Hi all,
More test
I had a class which creates 250 users, If I comment it, puppet is able
to run in 188 nodes at once with no problem.
so, definitely, scalability issues comes from code.
anyway, if someone could give me some tips on the OT will be appreciate.
Cheers,
Arnau
--~--~-~
Arnau Bria schrieb:
> Hi all,
>
> More test
>
> I had a class which creates 250 users, If I comment it, puppet is able
> to run in 188 nodes at once with no problem.
> so, definitely, scalability issues comes from code.
>
> anyway, if someone could give me some tips on the OT will be apprec
On Fri, 12 Jun 2009 12:34:25 +0200
David Schmitt wrote:
Hi David,
> Compiling the manifest is the single most resource-intensive
> operation the puppetmaster has to do.
Yep, I noticed that.
> Therefore it is not recommended to have all clients hit the server at
> the same time.
I know, but we
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Arnau Bria wrote:
>> For maximum performance you want to avoid any file-server operations
>> (file.source with puppet://, filebucket to server},
> I was thinking on implement an rsync in same puppet server and do a
> rsync Exec...
That'll work.
>>
I use the RPM's from plain EPEL for both RHEL 4 and 5, on both i386
and x86_64, and so far they've worked great.
If you are concerned about the packages, I would recommend setting up
a little test environment with a puppetmaster and a few clients, and
just test it out.
BR.
Rune Stensø
On 12 Jun,
Arnau Bria schrieb:
>> Therefore it is not recommended to have all clients hit the server at
>> the same time.
> I know, but we're trying to see if we can use puppet for immediate
> changes (I know puppet is not designed for that, for we'd like to test
> it).
Use the method I described in my ear
On Friday 12 June 2009 22:46:53 Arnau Bria wrote:
> > For maximum performance you want to avoid any file-server operations
> > (file.source with puppet://, filebucket to server},
>
> I was thinking on implement an rsync in same puppet server and do a
> rsync Exec...
Can't you have puppet:/// serve
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Robin Sheat wrote:
> On Friday 12 June 2009 22:46:53 Arnau Bria wrote:
>>> For maximum performance you want to avoid any file-server operations
>>> (file.source with puppet://, filebucket to server},
>> I was thinking on implement an rsync in same pup
Hi guys,
I was enlightened by the 'puppet+augeas modprobe.conf' post (thanks!)
and was going to put inittab under puppet/augeas management.
My goal is to add this one line for spawning a console on the serial
port if it doesn't exist:
co:12345:respawn:/sbin/agetty ttyS0 19200 vt100
However the
Hi
> What's the correct yum repo to use for installing Puppet & Facter on
> RHEL5 and Centos5?
>
> I used to get them from the dlutter-rhel5 repo but this seems to be
> massively out of date now - latest version of puppet-server in there is
> 0.24.5-1.el5 and facter 1.5.4-1.el5.
>
> In Epel I see
Trevor Hemsley wrote:
> What's the correct yum repo to use for installing Puppet & Facter on
> RHEL5 and Centos5?
>
> I used to get them from the dlutter-rhel5 repo but this seems to be
> massively out of date now - latest version of puppet-server in there
> is 0.24.5-1.el5 and facter 1.5.4-1.el5.
Todd Zullinger wrote:
> Generally, the most up to date packages are in epel-testing. If you
> enable epel-testing, you'll also want epel enabled. At the moment,
> that will get you the latest stable releases of puppet and facter
> (well, facter is at 1.5.5 instead of 1.5.6 because 1.5.6 doesn't
Trevor Hemsley wrote:
> I have EPEL itself enabled as one of my repos but epel-testing is
> generally disabled - enabled only when I need specific things from
> it.
That works.
> Currently epel-testing for RHEL5/CentOS5 appears to have a wrong
> package in it for facter (1.5.5-1.el4 vs .el5). I s
On Fri, 12 Jun 2009 21:02:03 +1000
James Turnbull wrote:
[...]
> >> don't use reports,
> > Wow... I use Gridview and we really like it... but maybe we could
> > live without it and adding some nagios sensors.
>
> What's GridView - is that a Puppet specific tool?
sorry, PuppetView ... I work in G
David Schmitt schrieb:
> Arnau Bria schrieb:
>>> Therefore it is not recommended to have all clients hit the server at
>>> the same time.
>> I know, but we're trying to see if we can use puppet for immediate
>> changes (I know puppet is not designed for that, for we'd like to test
>> it).
>
> Us
On Fri, 12 Jun 2009 13:07:25 +0200
David Schmitt wrote:
[...]
> > what diff are there between marshal and yaml? I mean, do I lose
> > information? reports ?
>
> AFAIK marshal is equivalent to yaml, only faster and not backwards
> compatible.
will try.
> >> don't use reports,
> > Wow... I use
On Fri, 12 Jun 2009 21:28:09 +1000
James Turnbull wrote:
> > Can't you have puppet:/// servers separate to the catalogue server?
> > Would that help?
>
> Yes - see http://reductivelabs.com/trac/puppet/wiki/PuppetScalability
>From my tests I saw that filserver is not a problem but users module i
Todd Zullinger wrote:
> Are you sure your repo configs are correct? I don't see a .el4 facter
> package in the epel-testing repos for EL5:
>
They are correct but they were not before. What I seem to have
discovered that I didn't know before is that `yum clean all` needs to be
told to --enabler
Trevor Hemsley wrote:
> They are correct but they were not before. What I seem to have
> discovered that I didn't know before is that `yum clean all` needs to be
> told to --enablerepo=epel-testing in order to clean the cache for that
> repo (yes, it's documented but only when you bother to look it
Hello
David Schmitt wrote :
> For maximum performance you want to avoid any file-server operations
> (file.source with puppet://, filebucket to server}, set the
I agree with you, mainly, fileserver is slow.
But few tests I made using 'file://' instead of 'puppet://' didn't speed
up my puppet
Arnau Bria a écrit :
> Hi all,
>
> More test
>
If your doing benchmarking with puppet
try:
# puppetd -t --summarize
Really helpful.
--
Aurelien Degremont
CEA
--~--~-~--~~~---~--~~
You received this message because you are subscribed to the Google Groups
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Aurelien Degremont wrote:
> Arnau Bria a écrit :
>> Hi all,
>>
>> More test
>>
>
> If your doing benchmarking with puppet
> try:
>
> # puppetd -t --summarize
>
Also puppet-test in the ext directory is useful too.
Regards
James Turnbull
- --
I think this could be nice to do. For clusters (where every node
should have the same entries) it would ease the handling.
If so, how do you manage it? Static file, template, augeas, other?
Would be interesting to know.
BR, Rene
--~--~-~--~~~---~--~~
You rece
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Rene wrote:
> I think this could be nice to do. For clusters (where every node
> should have the same entries) it would ease the handling.
>
> If so, how do you manage it? Static file, template, augeas, other?
>
> Would be interesting to know.
>
Rene wrote:
> I think this could be nice to do. For clusters (where every node
> should have the same entries) it would ease the handling.
>
> If so, how do you manage it? Static file, template, augeas, other?
On the Beowulf clusters where we run Puppet, we use a template. Since
we want over 15
2009/6/11 Jason Antman :
>
> I was wondering if anyone out there has done any work on other URI types
> for File "source" ?
>
> I have a few machines that need some local scripts copied over, which
> are stored in CVS or Subversion. It would be really helpful if, instead
> of having to check out t
2009/6/12 Aurelien Degremont :
>
> Arnau Bria a écrit :
>> Hi all,
>>
>> More test
>>
>
> If your doing benchmarking with puppet
> try:
>
> # puppetd -t --summarize
>
> Really helpful.
Interesting. Any reason why this option is not documented in puppetd
-h? I'm using 0.24.8 FWIW.
Cheers,
-
It would be great if I was able to refer to a variable in a class namespace
using only it's string value.
I guess this is a kind of eval-esque construct for those familiar with Perl
and Python.
Example source :-
^^^
$classname = 'foo'
class foo {
$bar = "baz"
no
On Fri, 2009-06-12 at 04:46 -0700, Hui wrote:
> Hi guys,
>
> I was enlightened by the 'puppet+augeas modprobe.conf' post (thanks!)
> and was going to put inittab under puppet/augeas management.
>
> My goal is to add this one line for spawning a console on the serial
> port if it doesn't exist:
>
Hi,
> > I look forward to all the people who manage module collections
> > putting in their ten cents of approach and design. Woot! Bring on
> > the merge fest!
>
> Yes, something I forgot to mention: I'm totally open to other people
> interested in this project to join, commit, and maintain t
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Hi all
Some of you might have gotten ticket emails about your ticket being
re-assigned from the Puppet "community" user. This doesn't mean we've
dropped your request or ticket. We've just decided that all tickets not
directly assigned to someone sh
On Fri, Jun 12, 2009 at 12:15 AM, David Schmitt wrote:
> Yes, something I forgot to mention: I'm totally open to other people
> interested in this project to join, commit, and maintain their modules.
>
>
> Regards, DavidS
I'll be glad to join this effort. How do I add myself to the project?
--Pa
34 matches
Mail list logo
