skip book previous and next navigation links
go up to top of book: HP OpenVMS Alpha Partitioning and Galaxy Guide HP OpenVMS Alpha Partitioning and Galaxy Guide
go to beginning of chapter: The Graphical Configuration Manager The Graphical Configuration Manager
go to previous page: Troubleshooting the GCM Server Troubleshooting the GCM Server
go to next page: Maintaining the GCM ServerMaintaining the GCM Server
end of book navigation links

Performance  



Because GCM is a client/server application, it is slower to respond to commands and configuration events than its native counterpart, the Galaxy Configuration Utility (GCU). While a Galaxy CPU assignment takes only a few milliseconds and is reflected immediately by the GCU, the GCM may take several seconds to update its display. For real-time response, run the GCU (and possibly multiple instances of the GCU if you have multiple hard partitions).

PC systems running Microsoft Windows require substantially more physical memory than may be present in a typical desktop PC. For best performance, your PC should have at least 128 MB (and preferably more). Disk space is seldom a concern.

The native OpenVMS GCM client does not perform quite as well as the Windows GCM client. On OpenVMS, user accounts must have substantial pagefile quota to run the GCM client. HP suggests that you set the PGFLQUO parameter to 350000 or more. (This is true of all Java applications on OpenVMS.)

Performance of GCM servers is typically a function of network behavior. GCM servers sit idle the majority of time, waking up every so often to issue a heartbeat transaction to other GCM servers or clients. The GCM server responds to configuration change events by reencoding the AlphaServer Configuration Tree structure in memory and then transmitting an XML- encoded representation to all active GCM servers and clients. This typically creates a burst of approximately 100 KB of network traffic. In an association that contains multiple GCM servers, with each server actively supporting GCM clients, the GCM servers must merge these bursts of data into a single configuration model, and then forward that model to each GCM client. The new model can require a megabyte or more.

To ensure optimal GCM performance, remember the following:


go to previous page: Troubleshooting the GCM Server Troubleshooting the GCM Server
go to next page: Maintaining the GCM ServerMaintaining the GCM Server