Puppet users - Nov 2013 - Continuous Integration Questions for Modules

Hi All,

Using Jenkins to perform CI as well as automated deployment of puppet 
modules to our master.

One thing that I am trying to figure out, is what is the best way when 
deploying the modules to have puppet not "error" out communication
wise if
nodes happen to check in while the deployment is ongoing.  I thought of 
using mcollective to stop puppet on all the nodes, then roll the new 
modules, then restart (puppet is in splay mode so should not overwhelm it).

Any thoughts/suggestions are appreciated!

Thanks!

Chuck

-- 
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+unsubscribe@googlegroups.com.
To view this discussion on the web visit
https://groups.google.com/d/msgid/puppet-users/0d6b3fe0-83ab-48c8-884d-27d7e320e0b7%40googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.

Have you seen such a problems or is just an assumption? I''ve worked on
a
few puppet implementations and changing puppet master manifests on the go 
has never been an issue.

I might be misunderstanding your question, can you develop more around your 
deployment process?

On Tuesday, November 19, 2013 9:01:27 PM UTC, gilbertc777
wrote:
>
> Hi All,
>
> Using Jenkins to perform CI as well as automated deployment of puppet 
> modules to our master.
>
> One thing that I am trying to figure out, is what is the best way when 
> deploying the modules to have puppet not "error" out
communication wise if
> nodes happen to check in while the deployment is ongoing.  I thought of 
> using mcollective to stop puppet on all the nodes, then roll the new 
> modules, then restart (puppet is in splay mode so should not overwhelm it).
>
> Any thoughts/suggestions are appreciated!
>
> Thanks!
>
> Chuck
>
-- 
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+unsubscribe@googlegroups.com.
To view this discussion on the web visit
https://groups.google.com/d/msgid/puppet-users/7c7871f0-186e-4f02-9e5a-ba497dd1de4a%40googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.

I''ve got to agree with JuanBrein, I''ve been using puppet since
0.24.5
and I don''t remember ever seeing a problem like this. How are you
deploying your modules?

Even if you have a "time consuming" deploy process (i.e. not just a
git
fetch and pull), and inconsistent state on disk is a valid concern, you
could easily bypass this by staging the deploy to a path and then either
renaming or switching symlinks to instantly activate the newly deployed
code/artifact.

-jantman

On 11/20/2013 06:39 AM, JuanBrein wrote:
> Have you seen such a problems or is just an assumption? I''ve
worked on
> a few puppet implementations and changing puppet master manifests on
> the go has never been an issue.
>
> I might be misunderstanding your question, can you develop more around
> your deployment process?
>
> On Tuesday, November 19, 2013 9:01:27 PM UTC, gilbertc777 wrote:
>
>     Hi All,
>
>     Using Jenkins to perform CI as well as automated deployment of
>     puppet modules to our master.
>
>     One thing that I am trying to figure out, is what is the best way
>     when deploying the modules to have puppet not "error" out
>     communication wise if nodes happen to check in while the
>     deployment is ongoing.  I thought of using mcollective to stop
>     puppet on all the nodes, then roll the new modules, then restart
>     (puppet is in splay mode so should not overwhelm it).
>
>     Any thoughts/suggestions are appreciated!
>
>     Thanks!
>
>     Chuck
>
> -- 
> 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+unsubscribe@googlegroups.com.
> To view this discussion on the web visit
>
https://groups.google.com/d/msgid/puppet-users/7c7871f0-186e-4f02-9e5a-ba497dd1de4a%40googlegroups.com.
> For more options, visit https://groups.google.com/groups/opt_out.
-- 
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+unsubscribe@googlegroups.com.
To view this discussion on the web visit
https://groups.google.com/d/msgid/puppet-users/528CAB19.8070900%40jasonantman.com.
For more options, visit https://groups.google.com/groups/opt_out.

+1 to Jason and Juan.
I have seen many installations but never had an issue with module updates
(eitherr by vcs or by os package).

- Martin

On 20 Nov 2013, at 13:29, Jason Antman <jason@jasonantman.com> wrote:
> I''ve got to agree with JuanBrein, I''ve been using puppet
since 0.24.5 and I don''t remember ever seeing a problem like this. How
are you deploying your modules?
> 
> Even if you have a "time consuming" deploy process (i.e. not just
a git fetch and pull), and inconsistent state on disk is a valid concern, you
could easily bypass this by staging the deploy to a path and then either
renaming or switching symlinks to instantly activate the newly deployed
code/artifact.
> 
> -jantman
> 
> On 11/20/2013 06:39 AM, JuanBrein wrote:
>> Have you seen such a problems or is just an assumption? I''ve
worked on a few puppet implementations and changing puppet master manifests on
the go has never been an issue.
>> 
>> I might be misunderstanding your question, can you develop more around
your deployment process?
>> 
>> On Tuesday, November 19, 2013 9:01:27 PM UTC, gilbertc777 wrote:
>> Hi All,
>> 
>> Using Jenkins to perform CI as well as automated deployment of puppet
modules to our master.
>> 
>> One thing that I am trying to figure out, is what is the best way when
deploying the modules to have puppet not "error" out communication
wise if nodes happen to check in while the deployment is ongoing.  I thought of
using mcollective to stop puppet on all the nodes, then roll the new modules,
then restart (puppet is in splay mode so should not overwhelm it).
>> 
>> Any thoughts/suggestions are appreciated!
>> 
>> Thanks!
>> 
>> Chuck
>> -- 
>> 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+unsubscribe@googlegroups.com.
>> To view this discussion on the web visit
https://groups.google.com/d/msgid/puppet-users/7c7871f0-186e-4f02-9e5a-ba497dd1de4a%40googlegroups.com.
>> For more options, visit https://groups.google.com/groups/opt_out.
> 
> 
> -- 
> 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+unsubscribe@googlegroups.com.
> To view this discussion on the web visit
https://groups.google.com/d/msgid/puppet-users/528CAB19.8070900%40jasonantman.com.
> For more options, visit https://groups.google.com/groups/opt_out.
-- 
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+unsubscribe@googlegroups.com.
To view this discussion on the web visit
https://groups.google.com/d/msgid/puppet-users/4CF0209B-038A-4881-B918-C4DAC1F387C6%40gmail.com.
For more options, visit https://groups.google.com/groups/opt_out.