Re: [Puppet Users] pip provider
Pip packages are case sensitive root@compute:~# pip freeze | grep -i xenapi XenAPI==1.2 On Mon, Jan 16, 2012 at 5:23 PM, Chris Blumentritt cblum...@gmail.com wrote: debug: Prefetching pip resources for packagedebug: Puppet::Type::Package::ProviderPip: Executing '/usr/bin/pip freeze'debug: Puppet::Type::Package::ProviderPip: Executing '/usr/bin/pip freeze'debug: Puppet::Type::Package::ProviderPip: Executing '/usr/bin/pip install -q xenapi'notice: /Stage[main]/Nova::Compute::Xenserver/Package[xenapi]/ensure: created root@compute:~# pip freeze | grep xenapi root@compute:~# pip install xenapi Requirement already satisfied (use --upgrade to upgrade): xenapi in /usr/local/lib/python2.6/dist-packages Cleaning up... On Mon, Jan 16, 2012 at 4:09 PM, Nigel Kersten ni...@puppetlabs.com wrote: On Mon, Jan 16, 2012 at 11:01 AM, Chris Blumentritt cblum...@gmail.com wrote: I am running puppet 2.7.9 and using the pip provider. The package gets/is installed but puppet installs it on every run. Do other people run into this. I tried search bugs and the group and have not seen any mention of it. Can you do a run with --debug on and paste the relevant parts Chris? -- Nigel Kersten Product Manager, Puppet Labs -- You received this message because you are subscribed to the Google Groups Puppet Users group. To post to this group, send email to puppet-users@googlegroups.com. To unsubscribe from this group, send email to puppet-users+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/puppet-users?hl=en. -- You received this message because you are subscribed to the Google Groups Puppet Users group. To post to this group, send email to puppet-users@googlegroups.com. To unsubscribe from this group, send email to puppet-users+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/puppet-users?hl=en.
[Puppet Users] pip provider
I am running puppet 2.7.9 and using the pip provider. The package gets/is installed but puppet installs it on every run. Do other people run into this. I tried search bugs and the group and have not seen any mention of it. Chris -- You received this message because you are subscribed to the Google Groups Puppet Users group. To post to this group, send email to puppet-users@googlegroups.com. To unsubscribe from this group, send email to puppet-users+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/puppet-users?hl=en.
Re: [Puppet Users] pip provider
That looks like exactly what is happening. I wonder if there is something special that has to happen in the pip package to show up when you do a pip freeze? Chris On Mon, Jan 16, 2012 at 4:10 PM, Dan Bode d...@puppetlabs.com wrote: I ran into the same issue when I tried to use the pip provider. Perhaps this is related? http://projects.puppetlabs.com/issues/7754 On Mon, Jan 16, 2012 at 11:01 AM, Chris Blumentritt cblum...@gmail.com wrote: I am running puppet 2.7.9 and using the pip provider. The package gets/is installed but puppet installs it on every run. Do other people run into this. I tried search bugs and the group and have not seen any mention of it. Chris -- You received this message because you are subscribed to the Google Groups Puppet Users group. To post to this group, send email to puppet-users@googlegroups.com. To unsubscribe from this group, send email to puppet-users+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/puppet-users?hl=en. -- You received this message because you are subscribed to the Google Groups Puppet Users group. To post to this group, send email to puppet-users@googlegroups.com. To unsubscribe from this group, send email to puppet-users+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/puppet-users?hl=en. -- You received this message because you are subscribed to the Google Groups Puppet Users group. To post to this group, send email to puppet-users@googlegroups.com. To unsubscribe from this group, send email to puppet-users+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/puppet-users?hl=en.
Re: [Puppet Users] pip provider
I have updated the bug with more info that Ben requested. On Mon, Jan 16, 2012 at 4:10 PM, Dan Bode d...@puppetlabs.com wrote: I ran into the same issue when I tried to use the pip provider. Perhaps this is related? http://projects.puppetlabs.com/issues/7754 On Mon, Jan 16, 2012 at 11:01 AM, Chris Blumentritt cblum...@gmail.com wrote: I am running puppet 2.7.9 and using the pip provider. The package gets/is installed but puppet installs it on every run. Do other people run into this. I tried search bugs and the group and have not seen any mention of it. Chris -- You received this message because you are subscribed to the Google Groups Puppet Users group. To post to this group, send email to puppet-users@googlegroups.com. To unsubscribe from this group, send email to puppet-users+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/puppet-users?hl=en. -- You received this message because you are subscribed to the Google Groups Puppet Users group. To post to this group, send email to puppet-users@googlegroups.com. To unsubscribe from this group, send email to puppet-users+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/puppet-users?hl=en. -- You received this message because you are subscribed to the Google Groups Puppet Users group. To post to this group, send email to puppet-users@googlegroups.com. To unsubscribe from this group, send email to puppet-users+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/puppet-users?hl=en.
Re: [Puppet Users] pip provider
debug: Prefetching pip resources for packagedebug: Puppet::Type::Package::ProviderPip: Executing '/usr/bin/pip freeze'debug: Puppet::Type::Package::ProviderPip: Executing '/usr/bin/pip freeze'debug: Puppet::Type::Package::ProviderPip: Executing '/usr/bin/pip install -q xenapi'notice: /Stage[main]/Nova::Compute::Xenserver/Package[xenapi]/ensure: created root@compute:~# pip freeze | grep xenapi root@compute:~# pip install xenapi Requirement already satisfied (use --upgrade to upgrade): xenapi in /usr/local/lib/python2.6/dist-packages Cleaning up... On Mon, Jan 16, 2012 at 4:09 PM, Nigel Kersten ni...@puppetlabs.com wrote: On Mon, Jan 16, 2012 at 11:01 AM, Chris Blumentritt cblum...@gmail.com wrote: I am running puppet 2.7.9 and using the pip provider. The package gets/is installed but puppet installs it on every run. Do other people run into this. I tried search bugs and the group and have not seen any mention of it. Can you do a run with --debug on and paste the relevant parts Chris? -- Nigel Kersten Product Manager, Puppet Labs -- You received this message because you are subscribed to the Google Groups Puppet Users group. To post to this group, send email to puppet-users@googlegroups.com. To unsubscribe from this group, send email to puppet-users+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/puppet-users?hl=en. -- You received this message because you are subscribed to the Google Groups Puppet Users group. To post to this group, send email to puppet-users@googlegroups.com. To unsubscribe from this group, send email to puppet-users+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/puppet-users?hl=en.
[Puppet Users] Adding node parameters to puppet dashboard
I would like to add parameters to nodes in puppet dashboard either from the command line or programmatically through an api. For adding classes and groups to a node as well as adding nodes to the dashboard I have been using the rake tasks. There is no rake task for adding parameters that I can find. Is there an api for the dashboard? Chris -- You received this message because you are subscribed to the Google Groups Puppet Users group. To post to this group, send email to puppet-users@googlegroups.com. To unsubscribe from this group, send email to puppet-users+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/puppet-users?hl=en.
Re: [Puppet Users] duplicate definition error
Basically, yes. It's all to do with how unqualified variables are resolved. Try class { '::passenger': } instead to qualify the class as being in the root scope. HTH Dan Thanks Dan that solved it. -- You received this message because you are subscribed to the Google Groups Puppet Users group. To post to this group, send email to puppet-users@googlegroups.com. To unsubscribe from this group, send email to puppet-users+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/puppet-users?hl=en.
[Puppet Users] duplicate definition error
I have an error that I do not understand. I get a duplicate definition and I am not sure why: http://pastie.org/3041278 if I change line 27 in that paste from class { 'passenger': } to include passenger, there is not an error but the passenger class (it is a module) is not evaluated. Does having module foo that contains a class, foo:bar::baz and then including another module named, baz cause this? I tested with puppet 2.7.9, 2.7.8 and 2.7.3 -- You received this message because you are subscribed to the Google Groups Puppet Users group. To post to this group, send email to puppet-users@googlegroups.com. To unsubscribe from this group, send email to puppet-users+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/puppet-users?hl=en.
[Puppet Users] Re: How long does the server's node cache stick around?
I don't know the answer to how long the cache hangs around or how to modify how long it will use the cache but I know that if you run puppet with the --test option it forces puppet to ignore the cache which is a lot less annoying then editing the yaml file. I suppose in the puppet.conf for your clients you could put ignorecache = true and it will always compile the manifest for you. Chris On Sun, Sep 27, 2009 at 12:50 PM, Tony Maro tonym...@gmail.com wrote: In my testing, puppet has never caught on to a new config or recipe to distribute unless I manually delete the node entry from the yaml/nodes directory. I couldn't find any documentation on how to clear the cache properly (I stumbled on that method) and I couldn't find any documentation on how long the cache is kept before it recompiles. All I saw were references to stale timestamps but no real details. I've waited as long as two days and it just doesn't see the change. If one day I decide I need to deploy a new required Gnome key to all of my Linux desktops, what's the poper method of making sure that the server realizes all the nodes need to be recompiled? I'm using LDAP for storing my node entries, if this makes a difference. All of my LDAP servers are doing live replication. --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups Puppet Users group. To post to this group, send email to puppet-users@googlegroups.com To unsubscribe from this group, send email to puppet-users+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/puppet-users?hl=en -~--~~~~--~~--~--~---
[Puppet Users] Re: cannot parse template when fact does not exist
On Fri, Sep 18, 2009 at 2:54 AM, David Schmitt da...@dasz.at wrote: Chris Blumentritt wrote: I receive the following error when trying to do the puppet below: Failed to parse template s_apache_site/site.conf.erb: Could not find value for 'ipaddress_eth0_0' As far as I can tell the ipaddress_eth0_0 fact has to exist on the client before it will parse. I was hoping that I could get the new ip address in place thus creating the fact so that when the apache site is created the information would be available. Maybe all facts are compiled at the when puppetd is run and there is no way to add facts during the run of puppetd? puppet sends all facts to the server at the beginning of each run to receive an updated configuration. The configuration is compiled (functions run, templates expanded) on the server before passing it on to the client. The latter only receives a list of resources which it dutifully applies. Therefore it is impossible to create facts on the fly: the client doesn't have enough information to re-compile its manifests. This is one of the few cases (enabling providers is another major one) where you have to defer some resources until the next run, where you have the facts available. The way I see around this is to first add the ip address with out calling the apache::site definition and once the ip address gets on there run puppet with apache::site definition. There has got to be better way, any suggestions If you really want to avoid the second run, you can refactor your manifest so you are not using a fact in you template, but directly the configuration information: node web1{ $secondary_ip = xxx network::ip { $secondary_ip: } apache::site { blah: ip = $secondary_ip } } How can I access the $secondary_ip variable from a class? I have the app_server class make the same site on more than one server and I do not think I can more than once in the node. Besides, defining the same site more than one time seems bad. Regards, DavidS --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups Puppet Users group. To post to this group, send email to puppet-users@googlegroups.com To unsubscribe from this group, send email to puppet-users+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/puppet-users?hl=en -~--~~~~--~~--~--~---
[Puppet Users] Re: cannot parse template when fact does not exist
On Mon, Sep 21, 2009 at 10:03 AM, David Schmitt da...@dasz.at wrote: Chris Blumentritt wrote: node web1{ $secondary_ip = xxx network::ip { $secondary_ip: } apache::site { blah: ip = $secondary_ip } } How can I access the $secondary_ip variable from a class? I have the app_server class make the same site on more than one server and I do not think I can more than once in the node. Besides, defining the same site more than one time seems bad. The variables propagate through the included classes. That way you can use $secodary_ip everywhere. I thought that would be the case but it does not seem to be working. I dropped some notices to print the ip address in the log of the puppet master and it is not printing anything. I had to take it a step further where I set another variable = to the secondary ip address so that I could use the same erb file for multiple projects. So I have it looking like this node app1 { $secondary_ip = x.x.x.x } class my_project::project_env { create_project { $project_fqdn: name = $project_name, project_network_interface = $secondary_ip } } snip from erb: VirtualHost %= project_network_interface -%:80 My apache conf ends up with this: VirtualHost :80 Regards, DavidS --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups Puppet Users group. To post to this group, send email to puppet-users@googlegroups.com To unsubscribe from this group, send email to puppet-users+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/puppet-users?hl=en -~--~~~~--~~--~--~---
[Puppet Users] Re: cannot parse template when fact does not exist
On Mon, Sep 21, 2009 at 12:19 PM, Chris Blumentritt cblum...@gmail.comwrote: On Mon, Sep 21, 2009 at 11:27 AM, David Schmitt da...@dasz.at wrote: Chris Blumentritt wrote: On Mon, Sep 21, 2009 at 10:03 AM, David Schmitt da...@dasz.at mailto:da...@dasz.at wrote: Chris Blumentritt wrote: node web1{ $secondary_ip = xxx network::ip { $secondary_ip: } apache::site { blah: ip = $secondary_ip } } How can I access the $secondary_ip variable from a class? I have the app_server class make the same site on more than one server and I do not think I can more than once in the node. Besides, defining the same site more than one time seems bad. The variables propagate through the included classes. That way you can use $secodary_ip everywhere. I thought that would be the case but it does not seem to be working. I dropped some notices to print the ip address in the log of the puppet master and it is not printing anything. I had to take it a step further where I set another variable = to the secondary ip address so that I could use the same erb file for multiple projects. So I have it looking like this node app1 { $secondary_ip = x.x.x.x } class my_project::project_env { create_project { $project_fqdn: name = $project_name, project_network_interface = $secondary_ip } } snip from erb: VirtualHost %= project_network_interface -%:80 My apache conf ends up with this: VirtualHost :80 Looks very good, but you also need to include my_project::project_env after defining the secondary_ip in the node. I have a separate class that includes the my_project::project_env for all of my app servers. I can verify that it is working since the apache conf file for site is there. I have been banging my head on a wall with this one. I have been dropping notice functions all over the place to figure out what is going on. In my create_project define I have added the following notices: notice(secondary ip address is ${secondary_ip}) notice(project network interface is ${project_network_interface}) and get the following output on the server: notice: Scope(S_project[my_project-env]): secondary ip address is x.x.x.x notice: Scope(S_project[my_project-env]): project network interface is I guess I am running into a scope issue or something. I have also tried to move the assignment of project_network_interface in front of the call to the create_project define. Does the behavior I am seeing make sense? I don't know what I can do different. Ok, the solution here is to define my ip address variables before I do my includes in the node. This is a bit annoying since I wanted to keep the definition of the variable together with my module that sets up the ip address on the server but at last it is working. Chris Regards, DavidS --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups Puppet Users group. To post to this group, send email to puppet-users@googlegroups.com To unsubscribe from this group, send email to puppet-users+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/puppet-users?hl=en -~--~~~~--~~--~--~---
[Puppet Users] cannot parse template when fact does not exist
I receive the following error when trying to do the puppet below: Failed to parse template s_apache_site/site.conf.erb: Could not find value for 'ipaddress_eth0_0' As far as I can tell the ipaddress_eth0_0 fact has to exist on the client before it will parse. I was hoping that I could get the new ip address in place thus creating the fact so that when the apache site is created the information would be available. Maybe all facts are compiled at the when puppetd is run and there is no way to add facts during the run of puppetd? The way I see around this is to first add the ip address with out calling the apache::site definition and once the ip address gets on there run puppet with apache::site definition. There has got to be better way, any suggestions The reason I am doing this is I will have an arbitrary number of servers load balancing a site and since I can keep track of what interface a site is going to use I can use that to build my apache file. -- node definition snippet: ip_address::new{ eth0:0: iface = eth0:0, ip = 192.168.0.84, netmask= 255.255.255.0, gateway= 192.168.0.1, iface_type = static, before = Apache2::Site[test.example.com] } app_server class snippet: $ipalias = ipaddress_eth0_0 apache2::site { test.example.com: name= test.example.com.conf, ensure = present, require = Apache2::Config[base], require = Ip_address::New[eth0:0], content = template(s_apache_site/site.conf.erb), notify= Exec[reload-apache2] } erb snippet: VirtualHost %= eval(ipalias) -%:80 --- Chris --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups Puppet Users group. To post to this group, send email to puppet-users@googlegroups.com To unsubscribe from this group, send email to puppet-users+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/puppet-users?hl=en -~--~~~~--~~--~--~---
[Puppet Users] Re: error starting puppetmaster after upgrade to 0.25.0
I overcame this issue by uninstalling all instances of puppet, re installing 0.25.0 and then changing into the gem directory and running install.rb by hand. I agree with the need for rubygems to distinguish between bin and sbin but I also need the executables :) On Mon, Sep 14, 2009 at 12:21 PM, Allan Marcus al...@lanl.gov wrote: see the thread New 2.5 gem does not create entries into /usr/sbin http://groups.google.com/group/puppet-users/browse_thread/thread/56ca66eee45bcb16?hl=en# --- Thanks, Allan Marcus 505-667-5666 On Sep 14, 2009, at 10:54 AM, Chris Blumentritt wrote: After an upgrade to 0.25.0 I get the following error: /usr/local/lib/site_ruby/1.8/rubygems.rb:383:in `bin_path': can't find executable puppetmasterd for puppet-0.25.0 (Gem::Exception) from /usr/bin/puppetmasterd:19 Uninstall just 0.25.0 and returning to .0.24.8 and it startup works. I also uninstalled all puppet gems and just tried to install 0.25.0 by itself. When I did this no executables were created in the /usr/bin directory for puppet. I am using hardy as the OS and installing via gem. Bug or is something whack in my environment? Chris --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups Puppet Users group. To post to this group, send email to puppet-users@googlegroups.com To unsubscribe from this group, send email to puppet-users+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/puppet-users?hl=en -~--~~~~--~~--~--~---
[Puppet Users] error starting puppetmaster after upgrade to 0.25.0
After an upgrade to 0.25.0 I get the following error: /usr/local/lib/site_ruby/1.8/rubygems.rb:383:in `bin_path': can't find executable puppetmasterd for puppet-0.25.0 (Gem::Exception) from /usr/bin/puppetmasterd:19 Uninstall just 0.25.0 and returning to .0.24.8 and it startup works. I also uninstalled all puppet gems and just tried to install 0.25.0 by itself. When I did this no executables were created in the /usr/bin directory for puppet. I am using hardy as the OS and installing via gem. Bug or is something whack in my environment? Chris --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups Puppet Users group. To post to this group, send email to puppet-users@googlegroups.com To unsubscribe from this group, send email to puppet-users+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/puppet-users?hl=en -~--~~~~--~~--~--~---
[Puppet Users] Re: Installing a lot of packages
On Sat, Sep 12, 2009 at 3:56 PM, ELTigre igalva...@gmail.com wrote: On Sep 12, 10:10 am, Chris Blumentritt cblum...@gmail.com wrote: Hi Chirs, I only use debian lenny on my servers. I just want to create a class with all package from APT to keep them up2date all the time with this configuration (ensure = latest) and whenever I create a new server it install all theses packages. The spirit of puppet is to install via the package resource. You could write a shell script to install using a text file as a data source then use the exec resource using onlyif to check to see if you need to run it. What How does onlyif fits here if I use exec? What should I check to prevent exec running every time puppet client get the catalog? You would have test to see if a directory that a packages does not exist. If it does not exist then install packages. This really is the wrong way to go about it though since this would be a huge pain to do for more than a few packages. I would do is create a separate class in its own manifest file and do some fancy searching and replacing (matching beginnings and endings of lines) with your list of packages. Since you have so many packages I doubt the resource will be more than one line. End up with something like: package { package_name: ensure = installed } Coul you explain me this, please? :-) This really is the best way to go have the package resource handle this. It will check to see if each package is installed without extra effort on your part. I am not sure what you want to explain about this, here is a link to the package documentation: http://reductivelabs.com/trac/puppet/wiki/TypeReference#package You might run into trouble here if you are using different distros of linux or linux and freebsd as the package names may be different. You would get the benefit of later being able to hit the file and change installed to latest if you want update a specific package though. I have never heard of a file being used as a source for a resource but it is an interesting thought. You might get away with an inline template to do that but that would be dirty: package { inline_template(%= filename.each_line.collect %): require = File[filename], ensure = installed I'll test it and let you know. thanks a lot. regards, israel } I have no idea if the above will work and I doubt the erb is correct since my ruby sucks. Chris On Sat, Sep 12, 2009 at 12:02 AM, ELTigre igalva...@gmail.com wrote: How can I install a lot (more than 800) of APT packages using an easy way in puppet? I mean, using some file.txt where I put there the packages I want to install and my package resource get the list . Is it possible? Or do I have to put all packages in the package resource? regards, Israel. --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups Puppet Users group. To post to this group, send email to puppet-users@googlegroups.com To unsubscribe from this group, send email to puppet-users+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/puppet-users?hl=en -~--~~~~--~~--~--~---
[Puppet Users] managing ip addresses
I am trying to manage ip address aliases and I cannot come up with a good strategy. I have three webhead servers that are balanced via a hardware load balancer. I have a define that will be setting up website config on each server and one piece I need to sort out is configuring ip addresses on the server and then configuring the apache virtual host with the corresponding ip address for the host. A sort of visual way to see what I am doing class site_name { ... } -- define setup_site { ...} The site_name class setups variables and the define does all the work based on the variables. I would like to setup a ip addresses that will be balanced for each site. And then do all of that with multiple sites. I fear the way to go about this is to set the ip address at the node and then use the fact ipaddress_ethx:y but this means I will have to have an interface to site mapping somewhere. I would rather just keep it all in puppet somehow. Any guidance is appreciated. Chris --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups Puppet Users group. To post to this group, send email to puppet-users@googlegroups.com To unsubscribe from this group, send email to puppet-users+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/puppet-users?hl=en -~--~~~~--~~--~--~---
[Puppet Users] Re: ssh_authorized_key completely ignoring require
I have run into this problem today trying to stand up some new servers. On Tue, Apr 21, 2009 at 11:44 PM, Andrew Shafer and...@reductivelabs.comwrote: Scott, Can you pastie the simplest code to reproduce and maybe attach the files created by --graph to see what the relationships look like. Is anyone else seeing a problem like this? On Tue, Apr 21, 2009 at 9:02 PM, Scott scott...@gmail.com wrote: Hi, so I'm running into a problem since upgrading to 0.24.8 where puppet is trying to create an authorized key for users that don't exist because it doesn't do the require ( require = /etc/passwd ) first. I've tried making the require a default parameter for ssh_autohrized_key (yes, in the same scope), I've tried making the passwd file a requirement for every ssh_authorized_key and I've tried to use before with the passwd resource ( before = Class [ users::ssh_keys ] ) and yet puppet insists on trying to create the key before doing any of the prerequisites. One other note, the ssh_authorized_key isn't always for the same person, so it's not a particular key that's causing the problem. Also, this was never a problem with 0.24.7. Cheers, Scott --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups Puppet Users group. To post to this group, send email to puppet-users@googlegroups.com To unsubscribe from this group, send email to puppet-users+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/puppet-users?hl=en -~--~~~~--~~--~--~---