The reload action sends a SIGHUP to the puppet master process, which is running as user puppet (privileges are dropped within puppet's own code), but was started as user root. In /usr/lib64/ruby/site_ruby/1.9.1/puppet/util/run_mode.rb (or equivalent path of course), the directory to look data in depends on the user that the process is started from: if root, it searches /var/lib/puppet, otherwise it searches ~/.puppet. What happens at this point is that the process is being executed as user puppet but still has HOME set to "/root", so puppet goes to search for the data in /root/.puppet — which it obviously can't access without root privileges, and thus will fail. We should remove reload() until this is fixed, probably upstream.
I think that's fine to remove til it's fixed for 3.1
Three in agreement (Theo said that he would do the same but didn't want to take the decision), so removed in 3.1.1-r1. I'll keep you guys posted about the upstream issue.
I'd like to keep this bug open til it's fixed (I know you said you'd keep us posted).
imho this bug should close since it is fixed given the current situation of the upstream package. if you are interested about any updates in the upstream bug you can add yourself there as a watcher (but again, I'm not willing to take the decision). Thanks for the investigation and the fix btw
worksforme
*** Bug 466286 has been marked as a duplicate of this bug. ***