Product Footprint

The Storage Automated Diagnostic Environment software was designed to have a very small footprint and to be invisible when not in use, it includes a cron and an on-demand http service used for browser, slave, and master communication.

The Storage Automated Diagnostic Environment software includes a cron that runs every five minutes. Every time the cron program starts, it verifies with the Storage Automated Diagnostic Environment configuration file if it is time to execute the agents.

The real agent frequency can be changed from agent to agent through the GUI. If, for example, the agent frequency was changed to 30 minutes, the cron will abort 5 times out of 6. This cron agent (/opt/SUNWstade/bin/rasagent) runs on both master and slave agents, and is a Perl program that can grow to approximately 15 Mb of memory. The Storage Automated Diagnostic Environment package does not include Perl, so a version of Perl must be present on the server for the Storage Automated Diagnostic Environment to work (Perl version 5.005 or higher). When running, the cron agent stores device-specific information in the /var/opt/SUNWstade/DATA directory. Its process size is not affected by the number of devices being monitored; once the monitoring of a device is completed, instrumentation data is stored on the disk and erased from memory.

Related Topics

The cron agent is only used to probe devices and generate events and does not provide access to the Storage Automated Diagnostic Environment GUI, Access is provided by an HTTP service, usually installed on port 7654 and 7443 (secure). This program, called /opt/SUNWstade/rashttp, is started from inetd and will stay in memory for as long as a user requires the GUI. Rashttp has a timeout period (default is 30 seconds), after which it exits. This minimizes the number of processes present on the servers. This HTTP service is also a Perl program, and its footprint is similar to the cron agent. It is used to answer HTTP requests coming from browsers or from slaves. Master and slaves use HTTP to share configuration information, topology information, and new events.