Puppet Server is intended to function as a drop-in replacement for the existing Apache/Passenger Puppet master stack. However, there are a handful of differences with Puppet Server due to changes in the underlying architecture.
This page details things that are intentionally different between the two applications. You may also be interested in the Known Issues page, where we’ve listed a handful of issues that we expect to fix in future releases.
Since the Apache/Passenger master runs under, well, Apache, the name of the service
that you use to start and stop the master is
on your platform). With Puppet Server, the service name is
to start and stop the service, you’ll use
puppetserver commands, such as
service puppetserver restart.
Puppet Server honors almost all settings in
puppet.conf and should pick them
up automatically. However, for some tasks, such as configuring the webserver or an external Certificate Authority, we have introduced new Puppet Server-specific configuration files and settings. These new files and settings are detailed on the Puppet Server Configuration page.
If you have server-side Ruby code in your modules, Puppet Server will run it via JRuby. Generally speaking, this only affects custom parser functions and report processors. For the vast majority of cases, this shouldn’t pose any problems, as JRuby is highly compatible with vanilla Ruby.
Ruby extension code in your modules needs to run under both Ruby 1.9 and Ruby 2.1. This is because Puppet Server runs Puppet functions and custom resource types under JRuby 1.7 (which is a Ruby 1.9-compatible interpreter), and the official
puppet-agent releases run custom facts and types/providers under MRI Ruby 2.1.
We isolate the Ruby load paths that are accessible to Puppet Server’s
JRuby interpreter, so that it doesn’t load any gems or other code that
you have installed on your system Ruby. If you want Puppet Server to load additional gems, use the Puppet Server-specific
gem command to install them. For more details on how Puppet Server interacts with gems, see the Puppet Server and Gems
Because Puppet Server runs on the JVM, it takes a bit longer than the Apache/Passenger stack to start and get ready to accept HTTP connections.
Overall, Puppet Server performance is significantly better than a Puppet master running on the Apache/Passenger stack, but the initial startup is definitely slower.
You can configure Puppet Server for use with an external CA instead of the internal Puppet CA. You’ll do this a little differently for Puppet Server than you would for an Apache/Passenger configuration. See the External CA Configuration page for more details.
See External SSL termination for details on how to get this working in Puppet Server.
The Passenger Puppet master used Puppet’s logging, which has 8 levels and
defaults to the ‘notice’ log level. Puppet Server uses logback, which has 5
levels and defaults to ‘info’. Puppet Server also combines the ‘notice’ and
‘info’ levels from Puppet and logs them at the ‘info’ level to logback.
Because of this mismatch in levels, log events that would not be logged by
default in Passenger will be logged in Puppet Server. In particular, anything
Puppet.info will be logged in Puppet Server and would not have
been under Passenger. See
Puppet Server Configuration Files: logback.xml
for more information, as well as instructions on how to set log levels in Puppet Server.