On Wed, Jan 21, 2009 at 3:59 PM, Luke Kanies <l...@madstop.com> wrote: > > On Jan 20, 2009, at 10:58 AM, Nigel Kersten wrote: > >> >> On Tue, Jan 20, 2009 at 7:27 AM, Ohad Levy <ohadl...@gmail.com> wrote: >>> O >>>> >>>> Something we've been meaning to have a look at is more related to >>>> how >>>> Puppet works with Facter than standalone facter I guess. >>>> >>>> A broken fact shouldn't break a Puppet run at all, and it is >>>> possible >>>> to do this at the moment. >>> >>> >>> I would agree with Nigel, I think he is referring to something like: >>> http://projects.reductivelabs.com/issues/show/1527 >> >> That's the one. The only time we've had to manually sort things out on >> puppet clients has been when we've accidentally pushed a broken fact >> to clients. > > Hmm, I think the only real way to fix this is to explicitly iterate > over all of the Facter facts, rather than using to_hash; that way > individual facts can fail without failing the whole process. > >> >> This is made somewhat worse by it being difficult to have facts part >> of the environment. We've set up a 'facts' module in each environment >> that is where clients get their facts from, but on first run you end >> up with a chicken or the egg problem where the server doesn't know >> what environment the client has, and even if you supply the >> environment on the command line, it's still ignored as far as the >> initial factsync goes. > > > Really? Is that filed? It's certainly a bug, if so. I know the > environment stuff isn't as thorough as it needs to be; it's obviously > caught me a bit by surprise at how hard it is to get every little piece.
I'll file it this week. We were attempting to work around it differently, but it hasn't worked out as well as we hoped. Which behavior do you consider to be the bug? The fact that the server doesn't take the environment into account until after the factsync process? > > BTW, looks like the RailsMachine guys have been working on some > extensions to both Puppet and Facter[1], and I hope to see a post from > Jesse Newland soon describing these. The Facter stuff is currently a > thin wrapper around the existing code, so it's completely backward > compatible but makes it cleaner and easier to use. > > 1 - > http://blog.railsmachine.com/articles/2009/01/16/moonshine-configuration-management-and-deployment/ > > -- > There are no such things as applied sciences, only applications of > science. -- Louis Pasteur > --------------------------------------------------------------------- > Luke Kanies | http://reductivelabs.com | http://madstop.com > > > > > -- Nigel Kersten Systems Administrator Tech Lead - MacOps --~--~---------~--~----~------------~-------~--~----~ You received this message because you are subscribed to the Google Groups "Puppet Developers" group. To post to this group, send email to puppet-dev@googlegroups.com To unsubscribe from this group, send email to puppet-dev+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/puppet-dev?hl=en -~----------~----~----~----~------~----~------~--~---
