Re: [Puppet Users] How to trace puppet execution to troubleshoot odd problems

[Jeff McCune]

On Tue, Oct 16, 2012 at 10:48 AM, HPUX_PUPPET <drid...@gmail.com> wrote:

What is the best way to get a full trace of what puppet is doing, with variable dumps and state, to troubleshoot problems like where a user's password keeps getting reset every time puppet runs even though the password hashes match? 

In these situations I try and run both the master and the agent with the following options turned on:  --verbose --debug --trace --evaltrace

 

I want to hack out a way to get HPUX working like we have our Linux environments but need to be able to track down right where a problem is coming in so I can adjust / override / submit changes to the code base.

With the --debug flag, providers will print out the exact command they're executing to modify the system.

--noop is also useful to see what commands Puppet _would_ execute without actually executing them.

-Jeff

[Jakov Sosic]

On 10/16/2012 08:49 PM, Jeff McCune wrote:

> With the --debug flag, providers will print out the exact command
> they're executing to modify the system.

What should a custom type/provider author put in provider code for
--debug to print the command out, or is it automatic?


--
Jakov Sosic
www.srce.unizg.hr

[Jeff McCune]

On Sat, Oct 20, 2012 at 7:09 PM, Jakov Sosic <jso...@srce.hr> wrote:

On 10/16/2012 08:49 PM, Jeff McCune wrote:

With the --debug flag, providers will print out the exact command
they're executing to modify the system.

What should a custom type/provider author put in provider code for --debug to print the command out, or is it automatic

It's automatic (but customizable) when the RAL synchronizes the "is" value to the "should" value.

See: https://github.com/puppetlabs/puppet/blob/master/lib/puppet/property.rb#L101-118

-Jeff