z/OS MVS Planning: Global Resource Serialization
Previous topic |
Next topic
|
Contents
|
Contact z/OS
|
Library
|
PDF
Contents (exploded view)
z/OS MVS Planning: Global Resource Serialization
SA23-1389-00
Abstract for MVS Planning: Global Resource Serialization
z/OS Version 2 Release 1 summary of changes
Global Resource Serialization
Introduction
Advantages of using global resource serialization
How global resource serialization works
Local and global resources
Using an API to serialize resources
ISGENQ macro
RESERVE macro
Potential problems when using RESERVE
Solving the RESERVE problems
Understanding the synchronous RESERVE feature
Issuing ISGQUERY and GQSCAN
ISGADMIN macro
Limiting global resource serialization requests
Methods of serializing global resources
The ring
The star
Star advantages
Which method to choose
Using IBM Health Checker for z/OS
Selecting the data
Data set naming conventions
Authorized qnames
Dynamic exits and RNL processing
Request processing sequence
RNL processing
How the RNL is scanned
ISGQUERY RNL search routine
SYSTEM inclusion RNL
SYSTEMS exclusion RNL
RESERVE conversion RNL
Excluding requests from RNL processing
RNL processing sequence
RESERVE conversion
Migrating from GRSRNL=EXCLUDE to a set of RNLs
Changing the RNL
RNL defaults
CICS
DAE
DB2
DFSMShsm
SMS
IMS
ISPF or ISPF/PDF
JES2
JES3
System logger
RACF
Tape volumes
Temporary data sets
TSO/E
SYS1.UADS and SYS1.BRODCAST
TSO/E user data sets
VIO journaling data set
VSAM data sets
Catalogs
RNL candidates
Defining the RNLs
Using the ENQ/RESERVE/DEQ monitor tool
Security considerations
Setting up the monitor
Monitor execution
Starting the monitor
Option 1
Option 2
Option 3
Option 4
Monitor control
Messages - abends - return codes
Filter facility
Reports
Logs
Temporary data sets
In-stream procedures variables
Report programs parameters
Region requirements
ENQ/RESERVE/DEQ restrictions
Resources used and guidelines
Monitor diagnostic information
Monitor physical output record
Monitor utilization hints
ENQ/DEQ/RESERVE analysis aid reports
Records layout
DSECT monitor physical and logical record
ISGAMEDM ISGAMED1 ISGAMVOL output record fields
EDSORTED output record fields
ISGAMCTM output record fields
Global Resource Serialization Star
Star processing
Global ENQ/DEQ processing overview
ISGQUERY and GQSCAN processing overview
Cross-system processing option
Processing system failures
Contention management
Contention notification
Planning a star complex
Decisions you need to make
Designing a star complex
Defining the sysplex couple data set for the star complex
Use of the sysplex couple data set
Dynamic RNL processing
The CFRM couple data set
Sizing the ISGLOCK structure
Getting the right structure size
Defining parmlib members for a star complex
Global resource serialization star definition parameter (GRSDEF)
Specifying the name of the system for GRSDEF parameters (MATCHSYS)
Specifying global resource serialization tracing options (CTRACE)
Activating synchronous reserve processing (SYNCHRES)
Collecting global resource serialization information (GRSQ)
Specifying the maximum ENQ requests (ENQMAXA, ENQMAXU)
GRSDEF considerations
Bringing up a star complex
Potential error scenarios
Error scenarios you can avoid
Initial processing environment
ISGLOCK rebuild processing
Coupling facility structure failure
Loss of connectivity to a coupling facility structure
Specifying global resource serialization tracing options (CTRACE)
Minimum options (MINOPS)
Operating the star complex
Operating the star complex
Building the complex
Configuration check
IPL
Normal operations in a sysplex
Displaying the status of systems in a star complex
Changing RNLs for a star complex
Getting a list of delaying jobs
Getting a list of suspended jobs
Responding to the ISG220D message
Cancelling the RNL changes or jobs
Rebuilding the ISGLOCK structure
Shutting down a coupling facility
Steps in converting to a star complex
Converting to a star complex
Using the SETGRS command to convert to a star complex
Conditions during dynamic conversion
Global Resource Serialization Ring
Ring processing
The RSA-message
Processing a request for a resource
Request processing without ring acceleration
Request processing with ring acceleration
Designing a ring complex
Designing a complex that matches a sysplex
Processing options in a sysplex
Residency time value (RESMIL)
Choosing a RESMIL value
Tolerance interval (TOLINT)
Ring acceleration (ACCELSYS)
Request processing
Recovery
ACCELSYS recommendations
Specifying global resource serialization tracing options (CTRACE)
Activating synchronous reserve processing (SYNCHRES)
Specifying the maximum ENQ requests (ENQMAXA, ENQMAXU)
Define the complex to MVS
Designing a mixed complex (sysplex does not match complex)
Choosing the link configuration
Link placement
Recovery
Level of connectivity
Alternate links
Backup considerations
Processing options in a mixed complex
System name (MATCHSYS)
CTC link device numbers (CTC)
Residency time value (RESMIL)
Configuration considerations
Workload considerations
RESMIL recommendations
Tolerance interval (TOLINT)
Ring acceleration (ACCELSYS)
Request processing
Recovery
ACCELSYS recommendations
Automatically rebuilding a disrupted ring (RESTART)
Example of automatic restart
Automatically rejoining the ring (REJOIN)
Recovery actions for a system failure with automatic restart
Specifying global resource serialization tracing options (CTRACE)
Minimum options (MINOPS)
Defining the complex to MVS
GRSCNFxx worksheet
Operating the ring complex
Operating a complex that matches a sysplex
Building the complex
Configuration check
IPL
Normal operations in a sysplex
Displaying ring status
Changing the RNLs for a ring
Operating a mixed complex (complex does not match sysplex)
Building the complex
Configuration check
IPL
Normal operations in a mixed complex
Displaying ring status
Quiescing a system
Example — Quiescing a System
Purging a system
Example — Purging an Active System
Example — Purging a Quiesced System
Restarting a system
Split rings
Example — Restarting a System
Controlling CTC links
Transitions to and from a mixed complex
Installing and tuning the complex
Installing the complex
Installing a new complex
Migrating an existing complex into a sysplex
Tuning the complex
Average response time
Tuning factors
Number of systems
Transmission rate
RSA-message size
Global resource request rate
RESMIL value
ACCELSYS value
Tuning process
Measurements
Using RMF
Using GTF
Global Resource Serialization Diagnosis
Diagnosing global resource serialization
Discriminating between system and application problems
Check if the complex is operating normally
Tuning the global resource serialization ring
Ring disruption recovery
ISG177E and ISG178E recovery
Global resource serialization ring rebuild
Checking XCF/XES connectivity and performance
ISGLOCK structure request processing
Checking for ENQ contention problems
Checking for latch contention problems
Using SMF 87 Records to identify global generic queue scan issuers
Measuring response time
Viewing system performance
Changing RESMIL values
Comparing response time
Steps for comparing global ENQ response time to RESERVE
Data set ENQ contention monitor
Installing the data set ENQ contention monitor
Assembling the data set ENQ contention monitor source module
Notes on the JCL to assemble the monitor module
Link-editing the data set ENQ contention monitor source module
Notes on JCL to Link-edit the ENQ contention monitor
Changes to SYS1.PROCLIB
Using the data set ENQ contention monitor
Starting the monitor
Stopping the monitor
Output
Recovery actions for global resource serialization
Loss of connectivity to the coupling facility
Recovery when some ring complex systems are in sysplex
Recovery when sysplex and ring complex contain same systems
Recovery actions for ring complex with some systems not in a sysplex
Purge a system from the ring complex before reIPL
Actions to purge a system before reIPL
Ring disruption in progress during IPL
Recovery actions for a ring disruption during IPL
Primary link failure
Recovery actions for a primary link failure
Alternate link failure
Recovery actions for an alternate link failure
System failure
Automatic restart example
Recovery Actions for a System Failure with Automatic Restart
Manual restart
Recovery Actions for a System Failure with Manual Restart
Restart of two-system complex
Recovery actions for a two-system complex if SYS1 fails
Reactivating a quiesced system
Recovery actions for reactivating quiesced systems
Trademarks
Copyright IBM Corporation 1990, 2014