--
You received this message because you are subscribed to the Google Groups "Puppet Users" group.
To view this discussion on the web visit https://groups.google.com/d/msg/puppet-users/-/Sm_12va2_aAJ.
To post to this group, send email to puppet...@googlegroups.com.
To unsubscribe from this group, send email to puppet-users...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/puppet-users?hl=en.
Sounds like you changed nsswitch.conf but haven't restarted puppet daemon. So puppet has no idea how to see those users. VFAQ ;-)
On May 29, 2012, at 2:06 AM, pattryn wrote:
I have this problem I can't figure out. I have a very simple define to deploy SSH keys to certain LDAP users:define sshkeys::install () {file{"/home/${title}":ensure => directory,owner => "${title}",group => "ldapgroup",}}which I then call using:sshkeys::install{["user1","user2","user3"]:}When I run puppet from the command line works just fine. It creates the home directories and and deploys the ssh keys (not shown in the example). However, puppet also runs as a deamon every 30 minutes and this time it always fails with "Could not find user XXXX". Can you think on any cause for this?I'm using puppet 2.7.5 on RH 5.6 with nss_ldapThanks--
You received this message because you are subscribed to the Google Groups "Puppet Users" group.
To view this discussion on the web visit https://groups.google.com/d/msg/puppet-users/-/Sm_12va2_aAJ.
To post to this group, send email to puppet...@googlegroups.com.
To unsubscribe from this group, send email to puppet-users+unsubscribe@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/puppet-users?hl=en.
On Wed, Jul 18, 2012 at 7:28 PM, pattryn <pat...@gmail.com> wrote:
> Nope.
>
> Curiously it works if I enable the nscd daemon but it's not very reliable
> and I would prefer not to use it.
>
nscd is kinda required to read your passwd database and such.
AFAIK,
puppet relies on it to read if a user/group exist <-- Can someone
re-confirm? :)
Sorry to be a thread-necromancer, but we're seeing the same problem on a tiny subset of our hosts. We saw the problem on a CentOS 5 x86_64 host, then it mysteriously went away, and now we're consistently seeing it on a CentOS 5 i386 host. As with pattryn, the resource applies successfully when run from the command line with puppet apply, but does not when Puppet is running with puppet agent -t. SELinux is in permissive mode. Things work fine when nscd is running, but like many CentOS 5 users, we'll never touch that thing in production.We saw this behavior originally on the 2.7 client, upgraded to the 3.0 client, and are still seeing the issue.Anyone have any ideas?