Puppet users - Jun 2007 - warning: fork terminates thread

I have recently notice puppet failing to restart failed services and 
found the following when i run puppetd in debug.

    debug: Service[ntpd](provider=redhat): Executing
    ''/etc/rc.d/init.d/ntpd status''
    /usr/lib64/site_ruby/1.8/puppet/util.rb:300: warning: fork
    terminates thread at
    /usr/lib64/site_ruby/1.8/puppet/external/event-loop/event-loop.rb:123
    debug: Service[ntpd](provider=redhat): ntpd status returned nil
    debug: Service[sshd](provider=redhat): Executing
    ''/etc/rc.d/init.d/sshd status''
    /usr/lib64/site_ruby/1.8/puppet/util.rb:300: warning: fork
    terminates thread at
    /usr/lib64/site_ruby/1.8/puppet/external/event-loop/event-loop.rb:123
    debug: Service[sshd](provider=redhat): sshd status returned nil

As a result a failed service is not restarted or a disabled service is 
not stopped. 

Tested with puppet 0.22.4 and 0.23.0 only on x86_64.

This error appears to occur with all service status checks on CentOS 4.4 
and 4.5 (ruby 1.8.1) and does not occur on CentOS 5 (ruby 1.8.5).

Any ideas?

Ben

Ben wrote:
> I have recently notice puppet failing to restart failed services and 
> found the following when i run puppetd in debug.
>
>     debug: Service[ntpd](provider=redhat): Executing
>     ''/etc/rc.d/init.d/ntpd status''
>     /usr/lib64/site_ruby/1.8/puppet/util.rb:300: warning: fork
>     terminates thread at
>     /usr/lib64/site_ruby/1.8/puppet/external/event-loop/event-loop.rb:12
further investigation revealed that the warning do not prevent puppet 
from restarting failed services instead bad exit status from an init 
script does.  instead of an exit status of 3 (or >0) naughty init 
scripts exit with 0 when a service is not running.

Ticket #684 patches the puppet redhat init scripts to behave correctly.

Ben