CoreCluster configuration

 

You should successfully go through the CoreCluster installation guide before starting this tutorial.

This document describes Computing Node configuration. Whole configuration is stored in /etc/corenode/ directory.

The most important changes, which you should do after CoreCluster installation are:

Any other modifications are not necessary, however may increase performance of your cloud. Usually after changing this files it is necessary to restart corecluster and/or uwsgi service. Safe way it to restart both.

config.py

DATABASES

Connection details for log and main databases.

CACHE_URL

Connection to Redis cache (by default this is localhost)

VNC_ADDRESS

IP address, which is used by nodes to listen for vnc connections. This is NOT the WebVnc port. To enable VNC redirections change it to 0.0.0.0

VNC_PORTS

Port pool used to redirect vnc connections. WebVnc uses the same ports + 10000

AUTOREGISTER_NODE

Defines if CoreCluster should accept registration requests from new nodes in network. Disable in production environments.

USER_QUOTA

Default values for user quotas. The quotas are stored in group definitions in database/administrator site

ENABLE_ALL_PERMISSIONS

Defines if CoreCluster should add all permissions to the default_role, created at the initial setup.

CREATE_ADDITIONAL_ROLES

Defines if CoreCluster should create new roles during the initial setup. This roles are: read_only, storage_role and default_role.

CHECK_STATES

If set to True, then CoreCluster's API checks if requested actions could be performed. If disabled, this checks are done only in Agents. Setting to True may result more frequent fails form the API.

MAX_UPLOAD_CHUNK_SIZE

Defines maximum size of upload to image, in bytes. Data is stored in cache. Too large values may result large memory consumption by cache.

NODE_SUSPEND_DURATION and NODE_WAKEUP_TIME

Defines how long node stays in suspend state and how long does it take to wake up node from suspend.

NODE_SELECT_ALGORITHM, STORAGE_SELECT_ALGORITHM, ID_GENERATOR and AUTH_METHOD

Define modules which are used to manage cluster's resources.

API_STATS

If set to True, all methods store information about call duration in databse. Enabling this may result increase of database usage.

INSTALLATION_ID

Random string used to identify logs in database (in case node and core have the same log database)

APPS

List of installed extensions. Add here any new extensions, like CoreTalk or CoreVpn. Each entry should point into app file in the extension module.

agents.py

WORKER_USER, WORKER_GROUP

User and group used to start agent processes

AGENTS

List of enabled agents. Add here any agents provided by additional extensions. Each entry defines type of tasks, module which contains agent thread and count of agents to start on system startup

MOUNT_NODES

Define if agents should mount all storages on all nodes in cluster. At least one agent should do this on start time.

REMOVE_TASKS_ON_DONE

If False, then done tasks are kept in database. It is useful in debugging. In production it may cause large usage of cache and slow task queues.

REMOVE_AGENTS_ON_DONE

If False, information about all agent processes (even closed) will be stored in database. Use for debugging only.

TASK_FETCH_INTERVAL

Defines how often agent should check for new tasks in queue

IGNORE_TASKS

List of task states, which could be skipped to execute new ones in queue

< Go back     Author: Maciej Nabozny Published: Sept. 23, 2016, 6:46 p.m.