-
Notifications
You must be signed in to change notification settings - Fork 22
Installation guide
Currently, Puli is only supported on Linux (32 or 64 bits). It was successfully tested with OpenSuse and RedHat.
- mysql-devel
- python-devel (for your version of python)
- MySQL-python
- python-sqlobject
- python-tornado
- python-psutil
Execute the following mysql commands to create the database and the dedicated user:
create database pulidb character set utf8;
create user puliuser identified by 'yourpasswd';
grant all privileges on pulidb.*
to 'puliuser'@'localhost'
identified by 'yourpasswd' with grant option;First you need to choose where will be your install dir will be. It can be either on a shared network or locally on each computer (recommended for the central dispatcher).
For this guide, the install dir will be /opt/puli, just make sure this dir is in your PYTHONPATH.
The program source folder are located in the repository OpenRenderManagement/Puli/src folder. From this path, you can copy octopus folder containing the code for the dispatcher and the worker to your install dir. Also, make sure to copy OpenRenderManagement/Puli/scripts.
In addition, you can copy the tools folder to your install dir. It contains a set of command line utilities to interact with the server.
Create the folders conf and logs in the install dir.
The settings are located in the file /opt/puli/octopus/dispatcher/settings.py.
Please check the following settings:
CONFDIR = /opt/puli/conf
LOGDIR = /opt/puli/logs
POOLS_BACKEND_TYPE = "file"
DB_ENABLE = True
DB_CLEAN_DATA = TrueTo launch the dispatcher, execute the following command in a shell:
python /opt/puli/scripts/dispatcherd.py -D -C
Upon the first execution, Puli will create the appropriate tables in the database.
For the following executions, you may set these parameters in the settings file:
POOLS_BACKEND_TYPE = "db"
DB_ENABLE = True
DB_CLEAN_DATA = FalseThese will tell Puli to init itself with the database that was previously created, and it will enable the persistence of the jobs.
If you want to erase the database in order to have a clean start, reset these parameters to the values in the installation section.
While dispatcher manages the joblist, the actual job is executed by the worker. You need at least one worker running. Worker installed on every computer that you want to use for renderfarm needs.
To get worker installed, copy the dirs in the same way as you did for dispatcher (see previous section).
python /opt/puli/scripts/workerd.py --debug --console --server 192.168.1.2 --serverport 8004
Make sure to replace server IP according to your case.