HP 3000 Manuals HP 3000 Manuals
Introducing GlancePlus/XL by Rex Backman--Application Support Division The Hewlett-Packard offering of real-time performance diagnostic tools for MPE XL based machines has strengthened with the introduction of GlancePlus/XL. This tool replaces Glance/XL as the HP diagnostic performance tool of choice. Users of Glance/XL will find the migration to GlancePlus/XL extremely simple. All original Glance/XL commands have been retained in GlancePlus/XL and the new added features retain the original look and feel of Glance/XL. Additionally, customers who have maintained Glance/XL software support will automatically receive the new GlancePlus/XL. By offering the original look and feel of Glance/XL and a new extended feature set, GlancePlus/XL provides a single solution for performance diagnostic situations. Users who need a simple look at the performance characterization of their machine will still find the original Glance/XL screen capabilities at hand. More detailed analysis can be performed by taking advantage of the new GlancePlus/XL screens and commands. Return on investments made in other Hewlett-Packard performance tools can be increased with GlancePlus/XL. Optional software linkages into the HP LaserRX/MPE Collector and the Software Performance Tuner (SPT/XL) provided in GlancePlus/XL allow for a new level of synergy between these performance tools. NEW SCREENS AND COMMANDS Four new screens are present in GlancePlus/XL that complement the original screens found in Glance/XL. These new screens provide a vehicle for isolating and identifying system performance anomalies. CPU Detail Screen The CPU Detail screen provides information on discrete global CPU utilization. Metrics are provided on the traditional components of CPU utilization such as Memory Manager usage, System Process utilization, ICS overhead plus other new MPE XL performance sensitive metrics. Included in these new metrics are items such as global switch rates, CQ SAQ or AST time, and CPU utilization by dispatcher queue. Disc Detail Screen System disc information is provided in the new Disc Detail screen. Individual disc drive metrics are provided that include values on disc drive utilization, physical read and write rates plus current and maximum queue depth values. A summary section provides information on logical disc drive activity and Memory Manager activity. Memory Detail Screen Physical memory activity is represented in the Memory Detail screen. Data can be found in this screen to help identify possible memory pressure situations. Metrics are provided that report global page fault rates, physical memory size and the system memory clock cycle delta. Filter Screen A powerful filtering command provides the mechanism for entry into the last of the new GlancePlus/XL screens, the Filter Screen. This screen, which is analogous to the Job/Session screen, provides a mechanism for the user of GlancePlus/XL to define a filter based on several criteria. Once defined, processes that match the selected criteria will be displayed in the Filter screen. Users can choose among several criteria items such as program name, user logon, or HP LaserRX/MPE Application name. Additionally, a boolean "And/Or" construct exists that allows for the processes to match all or just one of the selected criteria. The new "User" command adds more diagnostic functionality to the product by allowing GlancePlus/XL to identify what processes are accessing any permanent MPE XL file. Input to this command is a fully qualified file name and the resultant output is a list of the accessors of the file in question. MODIFIED SCREENS AND COMMANDS The Process Detail screen in GlancePlus/XL has been improved by the addition of new discrete process level performance metrics and a set of new process level subcommands. Data is now available at the process level on switch rates, Dispatcher status, and compatibility mode CPU utilization. New subcommands allow for on demand display of process open file information, process family tree data, and a process procedure trace. A new subcommand also provides a quick entry into the new Software Performance Tuner (SPT/XL) product that allows for a method to collect detailed process performance data concurrently from within GlancePlus/XL. The GlancePlus/XL Online Help System has incorporated new features to increase functionality. A base set of softkeys provide a way to easily navigate through the most often used sections of the Help System. General guidelines are now provided on data interpretation to help direct the user on possible system performance solutions. GLANCEPLUS/XL CUSTOMIZATION Two new features allow for increased user customization of GlancePlus/XL. The first is the ability to choose the "interesting process" option. GlancePlus/XL, as did Glance/XL, displays historical interesting process data in a push down stack display. In doing this, historical data is retained on the terminal memory with each interval separated by a blank line. This default operation can be changed to retain only the current interval's list of interesting processes. When set to this option by the use of the JCW GLANCECLEARSCREEN, screens that utilize the interesting process concept (Global, Job/Session and Filter) only retain the current interval's interesting process list. Four new configuration values have been added to further help in identifying interesting processes. GlancePlus/XL will display a process in certain screens only if it is interesting. Usually this means the process has consumed some CPU time, performed disc I/O, or has executed some terminal transactions. The new values used to further identify these interesting processes differ slightly in that they allow for processes to become interesting if the process is waiting for the CPU, waiting for a disc I/O to complete, waiting for memory, or is impeded. These new interesting reasons allow the user to identify processes that are waiting an excessive amount of time for critical system resources. Identifying critical processes that are waiting for resources, such as the CPU, allow for Operations personnel to quickly identify and proactively respond to possible performance issues.