Advanced OpenVMS Systems Management COURSE DESCRIPTION This course is designed to teach performance concepts relating to the OpenVMS system, and to use these concepts to develop a tuning methodology to interpret, monitor, and adjust performance mechanisms in OpenVMS. Practical uses of OpenVMS Internals, such as for crash dump analysis, process / resource issues, and general understanding of OpenVMS mechanisms, will also be shown. COURSE OBJECTIVES Upon completion of this course, a system performance analyst will be able to : understand fundamental performance concepts for memory management, CPU management, and I/O management in an OpenVMS system. The ability to use supplied monitors to interpret performance statistics is also included. Participants will also be able to do brief analysis of a systems crash, determine the reason for a process' MWAIT condition, and learn to use the System Dump Analyzer (SDA) utility. COURSE TOPICS ================= PERFORMANCE ====================== Performance Basics Factors affecting system performance Performance metrics Virtual system caching Memory Management in OpenVMS Memory layout in OpenVMS Process and balance slots Free and modified page lists Working set adjustment algorithms SWAPPER trimming Page and swapfile usage CPU Management in OpenVMS Software and hardware priorities concepts Scheduling in uni- and multi- CPU configurations Priority boosting Lock manager CPU use I/O Management in OpenVMS Buffered and direct I/O definitions Breakdown of disk I/O Measuring Disk and terminal I/O File system structure concepts File system caching VMScluster Performance Management in OpenVMS Additional software components in the VMScluster Cluster common system disks Distributed lock manager MSCP buffering Performance Modification Tools: MONITOR Overview of available tools Modification of Parameters SYSGEN AUTOGEN Performance Considerations/Modifications in: Memory management CPU management I/O management VMScluster management ========== PRACTICAL INTERNALS ================= Naming conventions used in the VMS executive System space locations VMS internal routines Data structure layouts and offset names Accessing VMS executive locations in DCL procedures Internals of the process P0 address space P1 address space Uses of P2 address space Process control block / Kernel thread block Process header Hardware process control block Job information block Systems components overview S0 /S1 / S2 address spaces Timer mechanisms SWAPPER, ERRFMT, OPCOM, and JOB_CONTROL processes Process creation System mechanisms Interrupt priority levels Interrupt handling Exception handling ASTs (Asynchronous system traps) Non-paged dynamic memory layout Mutexes and MWAIT states Analyzing system crashes (caused by software) SDA utility features Requirements to run SDA Bugcheck mechanism Bugcheck parameters Sample crash dump analysis (2) Using CLUE to analyze crash dump information Using the DELTA Debugger Scheduling Overview of scheduling states Scheduling states transitions MWAIT state Analyzing processes in MWAIT states COURSE DURATION This course requires five (5) days, 60 % lecture, and 40 % lab time. COURSE PREREQUISITES Participants are assumed to have background knowledge of VMS Systems Management (mainly for parameter information), and a working knowledge of DCL programming. |