
Making specifying the secure file optional (since it isn't used) means we should clarify that in the docs. Change-Id: Ie17cc08835861f3e6d4c9ca206797f2287769bd3
1.9 KiB
- title
-
Installation
Installation
Nodepool consists of a long-running daemon which uses ZooKeeper for coordination with Zuul.
External Requirements
ZooKeeper
Nodepool uses ZooKeeper to coordinate image builds with its separate image builder component. A single ZooKeeper instance running on the Nodepool server is fine. Larger installations may wish to use a multi-node ZooKeeper installation, in which case three nodes are usually recommended.
Nodepool only needs to be told how to contact the ZooKeeper cluster; it will automatically populate the ZNode structure as needed.
Statsd and Graphite
If you have a Graphite system with statsd
, Nodepool can
be configured to send information to it. Set the environment variable
STATSD_HOST
to the statsd
hostname (and
optionally STATSD_PORT
if this should be different to the
default 8125
) for the Nodepool daemon to enable this
support.
Install Nodepool
Install Nodepool prerequisites.
Nodepool requires Python 2.7 or newer.
RHEL 7 / CentOS 7:
yum install libffi libffi-devel @development python python-devel
You may install Nodepool directly from PyPI with pip:
pip install nodepool
Or install directly from a git checkout with:
pip install .
Configuration
Nodepool has one required configuration file, which defaults to
/etc/nodepool/nodepool.yaml
. This can be changed with the
-c
option. The Nodepool configuration file is described in
configuration
.
Although there is support for a secure file that is used to store nodepool configurations that contain sensitive data, this is currently not used, but may be in the future.
There is an optional logging configuration file, specified with the
-l
option. The logging configuration file is in the
standard python logging configuration
file format.