Puppet Community

Remember, with Puppet you want to define the end state (package is installed and service is running), not specify the logic.  Really, you would want to write this more as

```package { 'installer':
  ensure =&gt; 'latest',
  provider =&gt; 'choco'
}
service { 'installer':
  ensure =&gt; 'running',
  enable =&gt; 'true',
  require =&gt; Package['installer']
}```
Don't quote me on all of the exact syntax!  :slightly_smiling_face:

it's `provider =&gt; chocolatey,`  and I believe you'll need this module: <https://forge.puppet.com/modules/puppetlabs/chocolatey>

Lol yeah I caught that right away. Thank you!

I don't know if this helps but this sounds like a problem I ran into about a year ago.   Here's my notes:

### Problems installing Postgresql on Red Hat

Out of the box RHEL8 does not have Postgresql 11 available to it.  There is no module stream for psql11 so you must disable all of the existing postgresql module streams before running this role.  Yes, that sucks but as of this writing there is no fix yet.

Disable the built-in PostgreSQL module stream:

```
sudo dnf -qy module disable postgresql
```

A discussion of issues with installing Postgresql on RHEL derivitaves can be found [here](<https://www.symmcom.com/docs/how-tos/databases/how-to-install-postgresql-11-x-on-centos-7>).

Hi Will, we'll take a look at the documentation issue that tries to install postgres 11. As far as the second issue, we do not have a custom postgres rpm that gets around the problem outlined by Trevor.

Thanks for the update Mike. I'll see if I can compile myself and get it working then. Maybe I'll just submit a PR and see if I can poke Puppet folks to accept it

Welp, I tried Trevor's proposed fix of simply removing the section and it doesn't fix it either, just generates another different error. So... I'll hammer on it more later I suppose.

if you change it to `if (versioncmp($::operatingsystemrelease, '9') &gt;= 0) {` does that help?

it’s a puppet forge module, so i’m getting it from Puppetfile (hope my answer makes sense).  Should I make it local and change it?

Something is strange, you are calling with 2 arguments, not 3. Check if you are looking at the right version of source code.

```# more /etc/puppetlabs/code/environments/puppet7/modules/firewall/metadata.json
{
  "name": "puppetlabs-firewall",
  "version": "4.0.1",
  "author": "puppetlabs",
  "summary": "Manages Firewalls such as iptables",
  "license": "Apache-2.0",
  "source": "<https://github.com/puppetlabs/puppetlabs-firewall>",```

Unless versioncmp calls something internally with wrong number of args the reported error cannot be from the source you showed.

Hack the file and make it error a different way (call fail(”blah”) for example). If that does not happen then that is not the file being loaded.

Not sure if versioncmp always supported 3 args so could be newer code with older versioncmp (i.e too old puppet version).

Got a different error (matches the change I made)

you mentioned rhel8 host. what's the puppet version?

Puppet server - 7.9.5.  Puppet-agent - 7.23.0

<https://stackoverflow.com/questions/70861317/puppet-agent-upgrade-to-6-26-breaks-function-versioncmp> points to an issue with the puppetserver package actually.

&gt; This is most likely a dup of PUP-11434. Please see my comment on that ticket about restarting the puppetserver service after its local puppet-agent package is updated.


oh i did update puppet-agent after puppetserver was started.  WIll try it now.

Yes! That was it.  Got a different error message now, which is my issue now :slightly_smiling_face: THANK YOU!!!!

Seriously.  You rock <@U114TKH5M>.  I can’t even imagine how long it would’ve taken me to figure this one out :slightly_smiling_face:

glad to help! The puppetserver/puppet-agent thing has caught me out so many times

<@U11QGB2PN>++ Happy Birthday! :birthday:

Has nothing to do with hiera actually. It depends on the version of Ruby and what the YAML implementation that comes with it supports.

did you check the README.md from the puppetboard module?

I edited /puppetboard/manifest/init.pp
```class { 'puppetboard':
  python_version    =&gt; '3.6',
  manage_virtualenv =&gt; true,
  reports_count  =&gt; 40,
  default_environment =&gt; '*',
  manage_selinux =&gt; false,
}```
/puppetboard/manifests/apache/conf.pp
```class { 'puppetboard':
  python_version =&gt; '3.6',
}

# Access Puppetboard from <http://example.com/puppetboard|example.com/puppetboard>
class { 'puppetboard::apache::conf': }```
/puppetboard/manifests/apache/vhost.pp
```class { 'apache':
  default_vhost =&gt; false,
}

# Configure Puppetboard
class { 'puppetboard':
  python_version    =&gt; '3.6',
  manage_virtualenv =&gt; true,
}

# Access Puppetboard through <http://pboard.example.com|pboard.example.com>
class { 'puppetboard::apache::vhost':
  vhost_name =&gt; '<http://mypuppet.com|mypuppet.com>',
  port       =&gt; 80,
}```
what else have I missed?

think of them as libraries in other programming languages

if you've completely new to puppet I suggest to start with something simple to automate, to get to know the language

for example configuring DNS server, ntp server, rollout ssh keys

Hi <@U050S6NGFDM>! I can drop you some beginner guide in Russian if you’d like (guessing from your name :slightly_smiling_face:)

there is a Russian-speaking group in Telegram as well: <https://t.me/puppet_masters_ru>

here is the link: <https://habr.com/ru/company/avito/blog/507346/>

it’s a bit not well ordered from my point of view but good enough I’d say

<@U050S6NGFDM> just to ensure.. is `puppetboard/manifest/init.pp` the root of the puppetboard module? or is it your manifest in the production environment?

actually fell free to reask the question in telegram, we can proceed there if you’d prefer

corrected my mistake.
wrote a manifest to install puppetboard (path `./puppetlabs/code/environments/production/manifests/init.pp`
here is the manifest
```node default {
#default configurations
}
node '<http://puppetsrv.com|puppetsrv.com>' {
class { 'puppet board':
  python_version =&gt; '3.6',
  manage_virtualenv =&gt; true,
  reports_count =&gt; 40,
  offline_mode =&gt; true,
  default_environment =&gt; '*',
  manage_selinux =&gt; false,
}
class { 'apache':
  default_vhost =&gt; true,
}
# Access Puppetboard through <http://pboard.example.com|pboard.example.com>
class { 'puppetboard::apache::vhost':
  vhost_name =&gt; '<http://puppetsrv.com|puppetsrv.com>',
  port =&gt; 80,
}
class { 'puppetboard::apache::conf': }
}```
But I am getting an error
```Error: Could not retrieve catalog from remote server: Error 500 on SERVER: Server Error: Evaluation Error: Error while evaluating a Resource Statement, Duplicate declaration: Class[Apache::Mod::Wsgi] is already declared at (file: /etc /puppetlabs/code/environments/production/modules/puppetboard/manifests/apache/vhost.pp, line: 56); cannot redeclare (file: /etc/puppetlabs/code/environments/production/modules/puppetboard/manifests/apache/conf.pp, line: 44) (file: /etc/puppetlabs/code/environments/production/modules/puppetboard/ manifests/apache/conf.pp, line: 44, column: 3) on node <http://puppetsrv.com|puppetsrv.com>
Warning: Not using cache on failed catalog
Error: Could not retrieve catalog; skipping run```
What I did wrong. Thank you in advance

I'll be watching this thread - I tried about half dozen times, never did get it to work.  Finally ran out of time, so I went with Foreman.  Good luck!

Hmm.. I guess I can try to install it in a lab then.. didn’t thought it’s that hard

I gave up on Puppetboard as well.  It was easier to install manually last time I did it, which was admittedly quite some time ago.

FWIW, a quasi replacement for Puppetboard could be <https://github.com/skx/puppet-summary>

installed puppet-summary, configured, hit it with web browser and BAM!
Error: Binary was compiled with 'CGO_ENABLED=0', go-sqlite3 requires cgo to work. This is a stub

so it would appear one will need to build from source with the 'optional' C compiler built in, good times!

<https://go.dev/doc/install/source>

image.png

recompiled go from source with the CGO_ENABLED and lo and behold:

If `puppet-summary` is used in production, may want to think about pruning reports on a scheduled basis.

we're a tiny startup with ~ 10 hosts, we'll be under 100 for at least a year or two.  but yea.

One thing I'm stuck on is, I manually pulled in a bunch of reports &lt;hostname&gt;.yaml, before pointing my puppetserver at it, now my node summary thinks it has 43 hosts, instead of just 3, and I can't seem to figure out how to purge the old data from it.  I think it's using sqlite for it's db?  I've tried some stuff, but pretty limited in useful maintenance commands

Blow it away and recreate? I'll check what I've experienced in the past, in the meantime.

it may dwindle down over time, it was a test machine anyway, so I'll rebuild on my 'real' test box next, which is freebsd.  All I was really looking for really, just a way to view my environment, and honestly mostly for the 'suits'.  Mgmt loves this stuff.  LOL

Same for me _(in terms of getting visibility)_. Between `puppet-summary prune` and <https://github.com/skx/puppet-summary/issues/48#issuecomment-405802420> you might get it tuned properly.

I also use it as a quick diagnostic tool, since if you drill down deep enough you can look at changes, e.g.:

i don't even have 'sqlite3' installed anywhere.  where's yours?

what I am missing is a proper tool to analyse the DAG

reading the svg/dot files from the agent is... ugh. and they are not stored at a central place

Installed with sqlite rpm on CentOS Stream 8 located at `/bin/sqlite3`.

<@U1PC2RAH0>: do you have an example of the output you are looking for? Always looking for expanding my horizons... :-)

I am currently on my phone and cant create a demo. basically each puppet catalog is a DAG. if you run an agent with --graph it will create .dot files in /opt/puppetlabs/puppet/cache/state/graphs/ . having those in a web UI where you csn properly zoom in/out and search for nodes, that would be awesome

that enalbes people to see/debug classes/resources and their ordering. good to fix loops/improve performance

In the past I wanted ti hack something, then I noticed I suck at frontend design! :D

<@U01RH6J4QDC> -
PRAGMA writable_schema = 1;
DELETE FROM sqlite_master;
PRAGMA writable_schema = 0;
VACUUM;
PRAGMA integrity_check;

This removed my 'reports' table, but it returned after I bounced puppet-summary.

hmm, then a node checked in and it was 4, now it's 5...seems everytime an agent checks in, the 'unchanged' number keeps climbing.  I only have 3 nodes, strange behavior

<@U1PC2RAH0>: Visual Studio Code has this option "Open Node Graph to the Side" _(I couldn't get it to work but maybe you'll have better luck)_ 

2nd heading on this page has some more details <https://puppet-vscode.github.io/docs/features/puppet-commands/>

Yes, it was removed in Puppet Enterprise 2019.3.0 6.12.0 in favour of VS code:

Node graph removed:
- The node graph in the console has been removed due to infrequent use. The graph was used to view relationships between resources and classes within a node catalog. To generate a node graph now, use the Puppet VS Code extension.

Year I remember. I would say its the most frequent requested feature in <#C19V5U41Z|> :D

See the `abrader-gms` module at <https://forge.puppet.com/abrader/gms?src=slack&amp;channel=puppet>

Here's some notes I took when getting Puppetboard running in my environment:
# Puppetboard won't install properly with Python 3.8 so we have to use Python 3.6.
# Puppetboard config
puppetboard::default_environment: 'production'        # The default environment we want Puppetboard to present.
puppetboard::groups: puppetboard                      # Necessary so that the puppetboard user can access /srv/ files.
puppetboard::manage_git: true                         # Yes, install GIT packages.
puppetboard::manage_selinux: true                     # Ensure SELinux contexts get set.
puppetboard::manage_virtualenv: true                  # Install a Python 3.6 virtual environment in /srv (see $::puppetboard::python_version).
puppetboard::python_version: '3.6'                    # The default of Python 3.8 DOES NOT WORK for RHEL8 for some reason.

# Apache setup for the Puppetboard VHOST.
puppetboard::apache::vhost::vhost_name: "%{::fqdn}-puppetboard"  # The name of the virtual host file in /etc/httpd/conf.d
puppetboard::apache::vhost::port: 443                            # Of course, the listening port.
puppetboard::apache::vhost::ssl: true                            # Of course, the listening port.

puppetboard::puppetdb_host: "<http://puppetdb.host.example.com|puppetdb.host.example.com>"    # Where the PuppetDB server is that we're going to read from.

See the `puppet-puppetboard` module at <https://forge.puppet.com/puppet/puppetboard?src=slack&amp;channel=puppet>

That looks ok. The problem is probably on the lines before this statement

Sorry, ended up responding on the wrong post.

OK thank you; turns out you were right :)

Gives me  Could not parse for environment production: Syntax error at 'mysql::server'

```class { 'mysql::server':
        root_password           =&gt; 'password',
        remove_default_accounts =&gt; true,
        restart                 =&gt; true,
}```

Another Q,  given these two should only run on a server identified as a db server (currently done with a custom fact), how should I be including them? Seems the second one is an instance of a class?

class { 'mysql::server': root_password   ... }
mysql::db { 'mydb': ... }

Currently I'm doing a case $the_fact { include 'mysql::server' } type of thing

2nd one is defined resource of type ‘mysql::db’ (database I believe)..

ideally you should create a profile (`profile::mysql` maybe) which will do all the things (include mysql::server and create some databases), then include it into your database server role which will be included just for the database server :slightly_smiling_face:

i.e. `case $the_fact { 'database_server': { include 'role::database_server' } ... }`

i.e. `case $the_fact 'database_server' { include 'role::database_server' } ...`

So basically have a class that when included also includes the other classes needed, sounds like?

yes, I’d recommend you to read about roles and profiles