6c779a5f83
Based on the advice in this thread: http://groups.google.com/group/repo-discuss/browse_thread/thread/b91491c185295a71 Size the sshd threads (which are also used to handle https git requests) better. Based on current trends, we peak at about 50% cpu usage and 25% memory usage. Double the number of threads here to try to allow us to get closer to 100% cpu. (Corresponding increase in db connections.) Change-Id: Icde18233de01466b241ab28d38d2e98735108193 |
||
---|---|---|
manifests | ||
modules | ||
tools | ||
.gitreview | ||
README.md |
These are a set of puppet manifests and modules that are currently being used to manage some of the efforts of the OpenStack CI project. They are quite bare and crappy at the moment, but should grow soon.
Additionally, there is a script, make_puppet_lp.py which is used to generate a few lists of users from launchpad teams, to make management and population of user accounts on different types of servers easier.
There are currently two different entry points, the slave.pp and the server.pp manifest.
slave.pp is intended to be for jenkins slaves and adds all members of ~openstack-ci-admins
server.pp is intended as the base for other servers and adds members of ~openstack-admins
Puppet needs to be installed via gems, because we use the pip package provider for one of the packages and that is only in 2.7.
For instance:
/var/lib/gems/1.8/bin/puppet apply --modulepath=pwd
/modules manifests/slave.pp
or
/var/lib/gems/1.8/bin/puppet apply --modulepath=pwd
/modules manifests/server.pp