CCA Grid Node Installation › Grid Node Overview

Grid Node Overview

Grid processing is used to increase performance by distributing operational workloads to multiple grid nodes. CA Configuration Automation operations are grid-enabled so they can be divided into independent executable entities. These executable entities are distributed to available grid servers, grid nodes, and threads for execution. A server is capable of supporting multiple CCA Grid Nodes, each with multiple threads.

The CCA Server is installed with the master grid node. Unlike other grid nodes, the master grid node is configured to process a maximum of 10 simultaneous jobs. Additional grid nodes (that is, those installed by the grid node installation program) are configured to process a maximum of 20 jobs. The value for the master grid node is lower to conserve resources for other UI activities.

Grid nodes communicate with each other through a range of ports defined during installation (by default, ports 8065 through 8080). This port range is also specified on the Properties page in the Administration panel of the CCA Server UI. The ports are listed as the tcp.base.port and tcp.base.range properties.

When a grid node server starts, it creates a socket for grid-to-grid communication. CA Configuration Automation selects a grid node to execute tasks based on its availability. It is not possible to specify a specific grid node to execute tasks or to pair a grid node with a Management Profile.

CCA Grid Nodes are supported on Linux, UNIX, and Windows platforms and have their own installation programs. After installing a grid node and registering it with the CCA Server, grid processing is invisible to CA Configuration Automation users.

If you install additional grid nodes, locate them between managed servers and the CCA Database server. Grid nodes communicate with the CCA Database using a JDBC connection. Grid node servers communicate with CA Configuration Automation Agents installed on managed servers. Grid nodes can also use WMI, SSH, Telnet, or Web Service commands on managed servers when the CA Configuration Automation Agent is not used for component discovery or refresh operations.

Note:

• If you are considering performing component discovery on Windows managed servers without CA Configuration Automation Agents, at least one grid node or the CCA Server should be installed on Windows.
• When the Grid Node Service or CCA Server Service is installed inside the Windows domain, they should be configured to run with a domain user that has the Log On as a Service permission.
• Two grid node instances can run on a single server as long as the hardware can support them.