Hi all, I have managed to get puppet deployed on a small cluster of debian (virtual) machines. Very nice tool! I have a couple of questions, some of which I have found responses from google, but just don''t understand, others of which I have no idea at all. - is there anyway I can get emailed if a puppet client fails to update? One of my clients failed to run a script as apt-get needed a confirmation prompt in order to install something. The odd thing is that it worked on all the other nodes - strange. Anyway, I only found this out by noticing something it wasn''t being backed up (which required rsync which failed to install). It would have been nice to have been alerted by puppet - I use backuppc to backup each node so I have a puppet script which creates the backuppc user, adds the backuppc SSH key to the local backuppc~/.ssh/authorized_keys2. However, I still need to manually ssh to the new node from the backuppc (which isn''t controlled yet by puppet) in order for backuppc to get the node''s SSH key. Is there anyway I can automate this from the node, i.e. perform a ''ssh into backuppc'' ssh back to node and confirm the ''do you want to cache this key'' message''? This is all a bit backwards - what is the right thing to do here? - I monitor each node (nagios, groundworks, zenoss whatever). How can I create a custom automatically for each node? i.e. I want to add the new node to the nodes.pp list which inherits from a ''monitored_linux_machine'' node for example. Somehow, after some magic the monitoring server is aware of the new node. Same question for backuppc actually. For this last point I did consider a couple of alternatives: - have the puppet script create the configuration file and then copy it to the monitor using scp. - have the puppet script create the configuration file and then copy it to the puppetmaster''s configuration for the monitor Again, both of these feel a bit yicky and I am sure I have missed something. Anyway, that''s it - many many thanks for a great piece of software. I wish somebody would add the dashboard to a debian lenny repo ;) Col -- 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+unsubscribe@googlegroups.com. For more options, visit this group at http://groups.google.com/group/puppet-users?hl=en.
On Mon, Apr 12, 2010 at 5:24 PM, yatesco <colin.yates@gmail.com> wrote:> I have managed to get puppet deployed on a small cluster of debian > (virtual) machines. Very nice tool! > > welcome to the club :)> I have a couple of questions, some of which I have found responses > from google, but just don''t understand, others of which I have no idea > at all. > > - is there anyway I can get emailed if a puppet client fails to > update? One of my clients failed to run a script as apt-get needed a > confirmation prompt in order to install something. The odd thing is > that it worked on all the other nodes - strange. Anyway, I only found > this out by noticing something it wasn''t being backed up (which > required rsync which failed to install). It would have been nice to > have been alerted by puppet > > you can use the build in report called tagmail to send you changes orfailures a node has been reporting, or alternativily, you can set Foreman or Puppet Dashboard as a web gui. in this space, Foreman can also send you all kind of summary email reports (e.g. all hosts which failed in the last hour, day week etc).> - I use backuppc to backup each node so I have a puppet script which > creates the backuppc user, adds the backuppc SSH key to the local > backuppc~/.ssh/authorized_keys2. However, I still need to manually > ssh to the new node from the backuppc (which isn''t controlled yet by > puppet) in order for backuppc to get the node''s SSH key. Is there > anyway I can automate this from the node, i.e. perform a ''ssh into > backuppc'' ssh back to node and confirm the ''do you want to cache this > key'' message''? This is all a bit backwards - what is the right thing > to do here? > > you can trigger an exec which runs only if a certain condition has met (ornot)> - I monitor each node (nagios, groundworks, zenoss whatever). How can > I create a custom automatically for each node? i.e. I want to add > the new node to the nodes.pp list which inherits from a > ''monitored_linux_machine'' node for example. Somehow, after some magic > the monitoring server is aware of the new node. Same question for > backuppc actually. > > For this last point I did consider a couple of alternatives: > > - have the puppet script create the configuration file and then copy > it to the monitor using scp. > - have the puppet script create the configuration file and then copy > it to the puppetmaster''s configuration for the monitor > > Again, both of these feel a bit yicky and I am sure I have missed > something. >there are a few options in this space - 1. use storeconfigs, which enables your clients to say i need to be monitorord and your monitor server to know which clients he need to monitor (once a new client adds the network). 2. query with a function (or generate or similar) and parse the local data which is on your master, you could find all clients which has a fact or something like that 3. if 2 sounds like too much work, you can use Foreman query interface instead (http://theforeman.org/wiki/foreman/Query_Interface) Ohad -- 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+unsubscribe@googlegroups.com. For more options, visit this group at http://groups.google.com/group/puppet-users?hl=en.
Thanks Ohad - I didn''t know about foreman. Will have a play :) On Apr 12, 12:50 pm, Ohad Levy <ohadl...@gmail.com> wrote:> On Mon, Apr 12, 2010 at 5:24 PM, yatesco <colin.ya...@gmail.com> wrote: > > I have managed to get puppet deployed on a small cluster of debian > > (virtual) machines. Very nice tool! > > > welcome to the club :) > > I have a couple of questions, some of which I have found responses > > from google, but just don''t understand, others of which I have no idea > > at all. > > > - is there anyway I can get emailed if a puppet client fails to > > update? One of my clients failed to run a script as apt-get needed a > > confirmation prompt in order to install something. The odd thing is > > that it worked on all the other nodes - strange. Anyway, I only found > > this out by noticing something it wasn''t being backed up (which > > required rsync which failed to install). It would have been nice to > > have been alerted by puppet > > > you can use the build in report called tagmail to send you changes or > > failures a node has been reporting, or alternativily, you can set Foreman or > Puppet Dashboard as a web gui. > in this space, Foreman can also send you all kind of summary email reports > (e.g. all hosts which failed in the last hour, day week etc). > > > - I use backuppc to backup each node so I have a puppet script which > > creates the backuppc user, adds the backuppc SSH key to the local > > backuppc~/.ssh/authorized_keys2. However, I still need to manually > > ssh to the new node from the backuppc (which isn''t controlled yet by > > puppet) in order for backuppc to get the node''s SSH key. Is there > > anyway I can automate this from the node, i.e. perform a ''ssh into > > backuppc'' ssh back to node and confirm the ''do you want to cache this > > key'' message''? This is all a bit backwards - what is the right thing > > to do here? > > > you can trigger an exec which runs only if a certain condition has met (or > > not) > > > > > > > - I monitor each node (nagios, groundworks, zenoss whatever). How can > > I create a custom automatically for each node? i.e. I want to add > > the new node to the nodes.pp list which inherits from a > > ''monitored_linux_machine'' node for example. Somehow, after some magic > > the monitoring server is aware of the new node. Same question for > > backuppc actually. > > > For this last point I did consider a couple of alternatives: > > > - have the puppet script create the configuration file and then copy > > it to the monitor using scp. > > - have the puppet script create the configuration file and then copy > > it to the puppetmaster''s configuration for the monitor > > > Again, both of these feel a bit yicky and I am sure I have missed > > something. > > there are a few options in this space - > 1. use storeconfigs, which enables your clients to say i need to be > monitorord and your monitor server to know which clients he need to monitor > (once a new client adds the network). > 2. query with a function (or generate or similar) and parse the local data > which is on your master, you could find all clients which has a fact or > something like that > 3. if 2 sounds like too much work, you can use Foreman query interface > instead (http://theforeman.org/wiki/foreman/Query_Interface) > > Ohad-- 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+unsubscribe@googlegroups.com. For more options, visit this group at http://groups.google.com/group/puppet-users?hl=en.