Tarantella Administration Guide
> Arrays, servers and load balancing
> Introducing application server load balancing
Introducing application server load balancing
- Understand what application server load balancing is.
- Learn how Administrators manage application server load balancing.
- Learn about Tarantella Advanced Load Management.
- Learn the main steps for configuring application server load balancing.
|
What is application server load balancing?
Application server load balancing is concerned with:
- choosing an application server on which to run an application so that
users get the best performance; and
- distributing the launch of applications so that the application servers
achieve a similar relative workload.
How Administrators manage application server load balancing
Tarantella Administrators manage application server load balancing centrally
from the array by:
By default, Tarantella uses an algorithm which load balances
application servers by counting the number of application sessions each server
is hosting through Tarantella and then selecting the server with the fewest sessions.
You can also license Tarantella Advanced Load Management,
which provides additional algorithms for load balancing application servers
based on their true load when a user launches an application.
Using Tarantella Advanced Load Management
The Tarantella Advanced Load Management component provides
algorithms for load balancing application servers based on either the amount
of free memory or the amount of free CPU when the application is launched.
You can only load balance X applications, Windows applications and character
applications using these algorithms.
To use this feature, you must install a Tarantella Enhancement Module on every
application server. This installs a load balancing service which provides
Tarantella with 'real time' information about the application server's CPU/memory load.
It also helps Tarantella to detect whether an application server
is available, for example because it is being rebooted.
Overview of how the load balancing service works
The following is an overview of how the load balancing service works:
- Whenever the primary Tarantella server is started, it builds a list
of application servers that need to be load balanced. The list is updated
whenever a host is added to or removed from an application object.
- The primary Tarantella server contacts each of the application servers
that are to be load balanced and requests initial load information. It does
this by contacting the load balancing service which is listening on port 3579/tcp. Establishing
contact also confirms that the application server is available to run applications.
- The primary Tarantella server sends updates to the other array members
with capacity values for each of the algorithms and notifies them
which application servers are currently not available.
- The load balancing service sends regular updates to the primary
Tarantella server. The Tarantella server listens on
port 3579/udp for the updates. The updates take place even if the load does not change.
The absence of a regular update helps Tarantella to detect whether
a server is available to run applications.
- The primary Tarantella server sends updates to the other array members
with capacity values for each of the algorithms and notifies them
which application servers are currently not available.
The updates take place even if the load does not change.
- The primary or secondary Tarantella servers launch
applications on the basis of the load information they have received.
Note The load balancing service always sends application server load data
to the primary Tarantella server. If the primary is not available,
CPU or memory-based load balancing is not available and so the default
session-based load balancing is used instead.
Configuring application server load balancing
To use application server load balancing, you have to:
If you experience problems with CPU or memory-based load balancing, try
the load balancing troubleshooter.