HP NonStop G-Series - Manual (Page 10)

volume. Collectors are expected to read the port number from the configuration file. On the
NonStop, the communication takes place through IPC, with the collectors and worker writing
to the collator’s $RECEIVE. The names of the various processes are specified in a
configuration file and must be consistent. Collectors running in a server report data to the
Collator periodically. Collator maintains the latest performance data from each collector.
Worker: Worker in a managed server provides the HTTPS interface to components in the
CMS. Depending on the command received from the CMS, it interacts with the Collator to
provide the information back to the components in the CMS. The commands received from
the CMS components are embedded in the HTTP GET/POST request. The Worker interacts
with the collator and responds back to the requesting CMS component using XML messages.
On the Linux server, the worker is a PHP page that runs within the Apache Web Server on
that server, while on the NonStop server, the worker is a CGI interface managed by a
custom web server.
[Back to TOC]
3.2. Features
The HP NonStop Cluster Performance Essentials software provides comprehensive
performance monitoring and notification features. From a single window, it allows the
monitoring of the performance of a set of NonStop clusters. Using pre-defined and user-
defined rules, it determines and displays the performance health status of the NonStop
clusters and their servers for measured entities. The drill-down capability with the trending
graph feature allows users to identify the cause of a performance problem on an aberrant
node.
In addition to the monitoring capability, the Performance Essentials software provides a
mechanism to deliver notifications via email or SMS for pre-defined and user-defined rules.
Notifications are also logged to the Console to record any alerts.
[Back to TOC]
Performance Essentials HP SIM Client and NonStop & Linux Host User Guide – 544813 – 004
10