b55ed05a27
We'd like to be able to control sequencing of how and when puppet runs across our machines. Currently, it's just a set of agents that run kinda whenever they run. At times they hang and we don't know about it. Also, cross-server sequencing is impossible to achieve. Change the operation away from agents running on the machine as daemons, and instead ssh from the master to each machine. Change-Id: I76e41e63c6d0825e8735c484ba4580d545515e43 |
||
---|---|---|
.. | ||
dns.py | ||
launch-node.py | ||
README | ||
sshclient.py | ||
utils.py |
Create Server ============= Note that these instructions assume you're working from this directory on an updated local clone of the repository on the puppetmaster, and that your account is a member of the admin, puppet and salt groups for access to their respective keys:: sudo adduser $(whoami) admin sudo adduser $(whoami) puppet sudo adduser $(whoami) salt (Remember to log out and back into your shell if you add yourself to a group.) To launch a node in the OpenStack CI account (production servers):: . ~root/ci-launch/openstackci-rs-nova.sh export FLAVOR="8 GB Performance" export FQDN=servername.openstack.org sudo puppet cert generate $FQDN ./launch-node.py $FQDN --flavor "$FLAVOR" To launch a node in the OpenStack Jenkins account (slave nodes):: . ~root/ci-launch/openstackjenkins-rs-nova.sh export FQDN=slavename.slave.openstack.org nova image-list export IMAGE='Ubuntu 12.04 LTS (Precise Pangolin)' nova flavor-list export FLAVOR="8 GB Performance" sudo puppet cert generate $FQDN ./launch-node.py $FQDN --image "$IMAGE" --flavor "$FLAVOR" --salt The --salt option tells the script to automatically configure and enroll the server as a minion on the salt master. If you are launching a replacement server, you may skip the generate step and specify the name of an existing puppet cert (as long as the private key is on this host). The server name and cert names may be different and the latter can be specified with --cert if needed (older Jenkins slave types still use shared certs), but launch-node.py will assume they are the same unless specified. Manually add the hostname to DNS (the launch script does not do so automatically). Note that this example assumes you've already exported a relevant FQDN and sourced the appropriate API credentials above. When running outside the official OpenStack CI infrastructure, you will want to pass --server ci-puppetmaster.example.com otherwise the new node wil try to register with ci-puppetmaster.openstack.org - and fail hilariously. Add DNS Records =============== There are no scripts to automatically handle DNS at the moment due to a lack of library support for the new Rackspace Cloud DNS (with IPv6). However, the launch-node script will print the commands needed to be run to configure DNS for a newly launched server. To see the commands for an existing server, run: ./dns.py $FQDN