Puppet

This doesn’t seem to occur on any daemon that gets restarted by puppet (eg. when restarting httpd, but it does when restarting rsyncd).

First everything looks normal:

-bash-3.00# netstat -alp|grep rsync
tcp        0      0 *:rsync                     *:*                         LISTEN      9877/rsync         
tcp        0      0 *:rsync                     *:*                         LISTEN      9877/rsync

I run puppetd which will update my rsyncd.conf file and restart the rsync daemon:

-bash-3.00# puppetd --test --serve serv1 --server serv1
info: Starting handler for Runner
info: Starting server for Puppet version 0.22.0
info: Listening on port 8139
notice: Ignoring --listen on onetime run
info: Caching configuration at /var/lib/puppet/localconfig.yaml
notice: Starting configuration run
info: //serv1./systemimager/File[/etc/systemimager/rsyncd.conf]: Filebucketed to backup with sum 720e2bc25914a472146ccc05c1a3ec42
notice: /serv1./systemimager/File[/etc/systemimager/rsyncd.conf]/source: replacing from source puppet://serv2/overrides/etc/systemimager/rsyncd.conf
info: //serv1./systemimager/File[/etc/systemimager/rsyncd.conf]: Scheduling refresh of Service[systemimager-server-rsyncd]
notice: //serv1./systemimager/Service[systemimager-server-rsyncd]: Triggering 'refresh' from 1 dependencies
notice: Finished configuration run in 7.78 seconds

Then I try to run puppetd again:

-bash-3.00# puppetd --test --serve serv1 --server serv1
info: Starting handler for Runner
info: Starting server for Puppet version 0.22.0
Address already in use - bind(2)
/usr/lib/ruby/1.8/webrick/utils.rb:70:in @create_listeners'
/usr/lib/ruby/1.8/webrick/ssl.rb:87:in @listen'
/usr/lib/ruby/1.8/webrick/server.rb:63:in @initialize'
/usr/lib/ruby/1.8/webrick/httpserver.rb:24:in @initialize'
/usr/lib/site_ruby/1.8/puppet/server.rb:129:in @initialize'
/usr/sbin/puppetd:390:in @new'
/usr/sbin/puppetd:390

Now netstat gives the following output:

-bash-3.00# netstat -alp|grep rsync
tcp        0      0 *:rsync                     *:*                         LISTEN      9519/rsync         
tcp        0      0 *:8139                      *:*                         LISTEN      9519/rsync         
tcp        0      0 *:rsync                     *:*                         LISTEN      9519/rsync

Restarting the rsync daemon solves this issue:

-bash-3.00# /etc/init.d/systemimager-server-rsyncd restart
Stopping rsync daemon for systemimager: rsync.
Starting rsync daemon for systemimager: rsync.
-bash-3.00# netstat -alp|grep rsync
tcp        0      0 *:rsync                     *:*                         LISTEN      9877/rsync         
tcp        0      0 *:rsync                     *:*                         LISTEN      9877/rsync         

Something else: when trying to restart apache I get the following bug:

info: //ccmt-b./webserver/File[/etc/httpd/conf.d/repo.conf]: Scheduling refresh of Service"notice: //ccmt-b./webserver/Service[httpd":httpd]: Triggering 'refresh' from 1 dependencies
err: Found a bug: stack level too deep

However, the httpd daemon gets restarted though.

• Koen Vereeken