Running puppetd on CentOS5:

[root@puppettest ~]# uname -a
Linux puppettest.domain.com 2.6.18-92.1.22.el5xen #1 SMP Tue Dec 16 13:08:49 EST 2008 i686 i686 i386 GNU/Linux
[root@puppettest ~]# rpm -qa|grep ruby
ruby-1.8.5-5.el5_2.6
ruby-libs-1.8.5-5.el5_2.6
ruby-shadow-1.4.1-7.el5
ruby-augeas-0.2.0-1.el5
[root@puppettest ~]# rpm -qa|grep puppet
puppet-0.24.8-1.el5.1

When running "puppetd -d --no-daemonize", connections to puppetmaster time out:

debug: Calling fileserver.describe
err: Could not call fileserver.describe: #<Errno::ECONNRESET: Connection reset by peer>
err: //Node[generic_host]/std-ops/std-ops::centos/File[/etc/selinux/config]: Failed to retrieve current state of resource: Connection reset by peer Could not describe /std-ops/config: Connection reset by peer at /etc/puppet/modules/std-ops/manifests/centos.pp:13

After which previously loaded manifests are run and fileserver.describe called again. And then it fails again. This problem appeared yesterday after adding and modifying some puppet modules.

NOTE: puppetd and puppetmaster are on the same subnet behind the same switch, so this should not be a network problem.