Tag Info

Hot answers tagged

7

If they're still checking in, that's probably because the puppet master is not checking the CRL; they might not exist in the cert inventory on the master any more but they're still signed by the CA. Revoking trumps that but the revoke doesn't seem to be preventing their agent runs (verify they aren't just using cached catalogs with puppet agent --test). ...


5

Log into 300 different servers in a loop? for host in `cat allmypuppetboxes`; do ssh -o ConnectTimeout=5 root@$host '/etc/init.d/puppet stop ; rm -rf /var/lib/puppet/ssl /var/lib/puppet/ssl.expired ; puppet agent --server puppet.example.com --test --waitforcert 5' done


5

The most direct approach is running the agent with the --evaltrace flag like puppet agent --test --evaltrace This will give you a good idea of which resources are the most expensive. For your large files, you may want to experiment with different checksum alternatives.


3

An "out of sync" resource is one where the puppet agent detects that a change is needed and acts on it - so, for example, a file where the content or permissions doesn't match, or a service in the wrong state. This can differ from "changed" resources count for a run, in that the "changed" count is for successful changes, while the "out of sync" count will ...


3

If the exec resource has a dedicated stage, you can implement the desired behavior by having it subscribe to all other stages, e.g. exec { "pacify-rkhunter": ... subscribe => Stage['pre','main','aux'], refreshonly => true, }


3

According to this post, you can disconnect users using pkill -STOP -u USERNAME. You can create a resource definition something this in puppet: define kill_and_delete { exec { "killing $title": command => "pkill -STOP -u $title", onlyif => "grep '^$title' /etc/passwd", before => User[$title], } user { $title: ...


2

Between Puppet 2 and 3, config.ru (used by Passenger to run the Puppet master under Apache) needs updating as Puppet's initialisation changed. See the release notes here for the diff: Puppet 3 release notes: Puppet Master Rack Configuration Is Changed Foreman's installer sets this file up at /etc/puppet/rack/config.ru, but Debian users using ...


2

When something isn't builtin to file, I generally just to turn to an puppet exec declaration. By using find to execute whatever commands you mean by "manage files" you should be able to accomplish your goal. Find can exclude Named Pipes: -type c File is of type c: b block (buffered) special c character ...


2

There is nothing additional to install. On a node you configure /etc/puppet/device.conf with your switch information and then run puppet device on that node.


2

I concluded this to be a bug in apt. builddep notifies apt-get update instead of requiring it, which doesn't make much sense in my opinion. You want apt-get update to be executed before installing build dependencies. I patched apt to require apt-get update, which solves my dependency cycle. Submitted a pull request to apt to see what the devs say.


1

I see you do not have logs. I would stop puppetmaster on the master server then run it this way: puppet master --verbose --no-daemonize On the node server, stop puppet service and run it verbose with noop: puppet agent --test --noop This should give you some output that will hopefully help troubleshooting. My guess: Certificate issues. It usually is ...


1

you can use Fabric. Fabric is a Python (2.5-2.7) library and command-line tool for streamlining the use of SSH for application deployment or systems administration tasks.


1

mbarthelemy answer got me halfway there, but in the end, I had to tweak a little more: In my VagrantFile, I added this to the mapping to get it to work: :linux__nfs_options => ["no_root_squash"], :map_uid => 0, :map_gid => 0


1

The incl parameter takes the path to the file in the filesystem, not in the Augeas tree, so you must remove /files from the incl parameter. Because you put /files in the incl parameter, Augeas cannot find the file to parse, and ins tries to insert before a non-existing path, which is an error.


1

is there any reason to not use Puppet and The Foreman for workstations? No It is insane to manage about 1000 worksations or more with this configuration? No


1

As long as the master is operating normally, the agent will update the timestamp on $vardir/state/state.yaml on every run. A failing master (including compiler errors etc.) will lead to an outdated state.yaml file. You can use its freshness to determine whether your master is up and well.


1

Puppet makes some assumptions about the permissions of some files in /etc/puppet (and the directory itself, I believe). As such, you likely will not be able to make that location writeable to your regular account. This is usually not a good choice anyway. If you want to work in /etc/puppet directly, you should work as root - sudo is an ill fit here. Try ...


1

The issue has been resolved by adding following statement in unless => "/sbin/swapon -s | /bin/grep file > /dev/null"


1

The module has a syntax error so Foreman's proxy can't parse it to report the classes inside. Run Puppet's validator to see this: $ find . -name *.pp -exec puppet parser validate {} + Error: Could not parse for environment production: All resource specifications require names; expected '%s' at andlinger-java-1.0.1/manifests/init.pp:17 The line in ...


1

yum grouplist --verbose 2>/dev/null | sed -n "s/.*(\([^)]\+-support\)).*/'\1': ensure => absent;/p" Using double quotes for the sed script since you can't embed single quotes in a single quoted string. With your sample input: echo " Maori Support (maori-support) [mi] Mongolian Support (mongolian-support) [mn] Mynamar (Burmese) Support ...


1

You terminate a session by killing its parent process, called the session leader. Find out which process it is with: ps -dN|grep pts/3 or As mentioned by @Nitz pkill -9 -u username And for reference check this link: How can I force other users to log out?



Only top voted, non community-wiki answers of a minimum length are eligible