that's even worse

probably a bad time to ask this since it is EOD for most, but: How are people handling keeping deployed environments in sync with multiple compile servers? I'm in the process of migrating to puppetserver 7 with an external CA and arbitrary compile nodes. Currently with my single puppetserver, a git push triggers an r10k deploy on the server, but it seems kinda flaky to attempt to do that on multiple (arbitrary) servers that may scale-in/out at any time. I've considered NFS and other shared filesystems. The puppet servers will be in ec2, likely across regions and definitely across availability zones and maybe also in our datacenter, too, making NFS problematic.

This message was deleted.

The two strategies boil down to “push” or “pull” models. • push: On each update to a branch, your git servers use a hook to trigger a listening service that tells the Puppet servers to run

r10k deploy

. • pull: your puppetservers run

r10k deploy

on a timer and constantly attempt to pull. Maybe every 1-10 minutes.

The two strategies boil down to “push” or “pull” models. • push: On each update to a branch, your git servers use a hook to trigger a listening service that tells the Puppet servers to run

r10k deploy

. • pull: your puppetservers run

r10k deploy

on a timer and constantly attempt to pull. Maybe every 1-10 minutes.

@vchepkov There is another way:

$squash_ensure = $squash_threshold.then |$x| { 'present' }.lest || { 'absent' }

🙂 not a selector too though 🙂

sporadically. i have this error displayed. Once displayed it makes the puppet server unusable. i had to force service restart.

Error: Could not retrieve catalog from remote server: Error 500 on SERVER: Server Error: Evaluation Error: Error while evaluating a Resource Statement, Evaluation Error: Error while evaluating a Resource Statement, Could not autoload puppet/type/stage: Attempt to redefine entity '<http://puppet.com/2016.1/runtime/type/stage>'. Originally set at file:///opt/puppetlabs/puppet/lib/ruby/vendor_ruby/puppet/type/stage.rb?line=1. (file: /etc/puppetlabs/puppet/g10k/production/modules/legacy/modules/profile/manifests/common1/stages.pp, line: 2, column: 3) on node <http://01.internet.net|01.internet.net>

Do you know how to troubleshoot this error?

# content of  modules/profile/manifests/common1/stages.pp
class profile::common1::stages {
  stage { [firstpkgmgmt, secondkerberize, last]: }
  Stage[firstpkgmgmt] -> Stage[secondkerberize] -> Stage[main] -> Stage[last]

} am running puppet v6

sporadically. i have this error displayed. Once displayed it makes the puppet server unusable. i had to force service restart.

Error: Could not retrieve catalog from remote server: Error 500 on SERVER: Server Error: Evaluation Error: Error while evaluating a Resource Statement, Evaluation Error: Error while evaluating a Resource Statement, Could not autoload puppet/type/stage: Attempt to redefine entity '<http://puppet.com/2016.1/runtime/type/stage>'. Originally set at file:///opt/puppetlabs/puppet/lib/ruby/vendor_ruby/puppet/type/stage.rb?line=1. (file: /etc/puppetlabs/puppet/g10k/production/modules/legacy/modules/profile/manifests/common1/stages.pp, line: 2, column: 3) on node <http://01.internet.net|01.internet.net>

Do you know how to troubleshoot this error?

# content of  modules/profile/manifests/common1/stages.pp
class profile::common1::stages {
  stage { [firstpkgmgmt, secondkerberize, last]: }
  Stage[firstpkgmgmt] -> Stage[secondkerberize] -> Stage[main] -> Stage[last]

} am running puppet v6

This message was deleted.

Hello again , Puppet beginner question :- Is it possible to run an exec only on a file location ? For example I would like to execute the below :-

exec { 'yum-clean-expire-cache':
    command     => '/usr/bin/yum clean expire-cache',
}

only on the the location where my yum repo config files are located . i.e.

/etc/yum.repos.d/

Please advise . Thanks

Hello again , Puppet beginner question :- Is it possible to run an exec only on a file location ? For example I would like to execute the below :-

exec { 'yum-clean-expire-cache':
    command     => '/usr/bin/yum clean expire-cache',
}

only on the the location where my yum repo config files are located . i.e.

/etc/yum.repos.d/

Wondering if this could be a good way to do it ?

exec { 'yum-clean-expire-cache':
    command => '/usr/bin/yum clean expire-cache',
    cwd     => '/etc/yum.repos.d/',
  }

Please advise . Thanks

message has been deleted

However , is it possible to get some pointers on this on how to create a Puppet task? along with also trying to touch a file . Would be a helpful learning .

This message was deleted.

Hi Team, I am using puppet server and agent version 6.12.0 on my setup, want to upgrade to 7.x. what is best way to do same. When I am trying to upgrade puppet server to 7.x it says puppet agent 6.16 or later is required. Shall I upgrade agent to 6.21.1 and then upgrade puppet server rpm to 7.x. Do we need to restart puppetserver service after upgrading puppet agent? I am following this guide https://puppet.com/docs/puppet/7/upgrading-from-puppet6-to-puppet7.html

(Sorry for the repost, meant to alse send to channel) @natemccurdy, thanks fo r the reply but unfortunately, it doesn't really answer the question. The question is not how to get updates to a single server, but how are people actually keeping multiple compile servers in sync with each other. Pull could potentially work, but a push would have to be parallel and is fraught with danger if one push failed. For any new server scaled-in, an initial pull-type sync could be triggered before the service is started. That's the easy part and what I'm doing now with a single server setup in an autoscaling group (read: terminate/launch on dead instance). In this case, since a cold backup is "good enough", I simply rsync the ssl dir regularly and trigger an rsync pull and r10k pull on our (currently 446) environments (it's a big infrastructure) and only after that, does the service start. But in the current scenario, I don't have to maintain sync on the live server. We have potentially dozens, if not hundreds, of commits on many branches happening every couple minutes. And while we often do instance refreshes to push out releases, I can't have some hosts getting different catalogs than others in the same cluster. Eventually consistent doesn't work for releases to prod.

Hi Team, I am using puppet server and agent version 6.12.0 on my setup, want to upgrade to 7.x. what is best way to do same. When I am trying to upgrade puppet server to 7.x it says puppet agent 6.16 or later is required. Shall I upgrade agent to 6.21.1 and then upgrade puppet server rpm to 7.x. I am following this guide https://puppet.com/docs/puppet/7/upgrading-from-puppet6-to-puppet7.html

Is bolt an enterprise-only feature?

No. Bolt is open source and free to use.

The enterprise only feature that builds on top of Bolt is the Puppet Orchestrator.

Thanks for info @bastelfreak, Can you please help me to find best way to migrate from 6.12.0 to 7.x

• I do have test environment. I am preferring to upgrade my current environment from 6.12 to 7.x. I can create another environment to test this upgrade approach. If that works then I can do the same changes to current environment. • Also I am not able to upgrade puppet server rpm from 6.12 to 7.x its says puppet-agent rpm 6.18 required. so do I need to upgrade puppet-agent rpm to 6.18/6.21.1 first then try upgrade to 7.x?

Is puppet on containers still being worked on (e.g. puppet ware) or did that get back burnered? I know there are more efforts to containerize puppet products like CD4PE but didn't know how that was going with the core stuff

This message was deleted.

I saw that in https://puppet.com/docs/puppet/7/lang_template_epp.html too, pretty sure it’s just a copy-paste error in the docs, from the ERB docs.

Hello again , Do you have any pointers as to how I could only run this exec when there is a change in the repo file ?

The target that I want to achieve is that - i would want to update the yum cache only when there is a change in the repo. So the issue that I am facing now is that when the repo configuration is broken and the latest version is used - then my facts file is being targeted as empty and it does not restore itself since the yum cache is not updated.

ERB and EPP handle using variables differently

once you get it working for one, wrap it inside a loop where you feed in the list of repo names

This message was deleted.