Part 1. Cics recovery and restart concepts | IBM SC34-7012-01

Part 1. CICS recovery and restart concepts

It is very important that a transaction processing system such as CICS can restart and recover following a failure. This section describes some of the basic concepts of the recovery and restart facilities provided by CICS.

© Copyright IBM Corp. 1982, 2010

1

Image 13

IBM SC34-7012-01 manual Part 1. Cics recovery and restart concepts

Contents

Recovery and Restart Guide Page Recovery and Restart Guide Copyright IBM Corporation 1982 Contents Restart Part 3. Implementing recovery Programming for recovery 141 Part 4. Appendixes 241 Disaster recovery 223 Preface How to use this bookWhat this book is about Who should read this book Viii Cics TS for z/OS 4.1 Recovery and Restart Guide Copyright IBM Corp Cics TS for z/OS 4.1 Recovery and Restart Guide Part 1. Cics recovery and restart concepts Cics TS for z/OS 4.1 Recovery and Restart Guide Maintaining the integrity of data Recovery and restart facilitiesLogging changes Role of Cics Minimizing the effect of failures Cics backward recovery backout Recoverable resources Emergency restart backout Dynamic transaction backout Forward recovery of Cics data sets Cics forward recovery Cics recovery processing following a communication failure Failures that require Cics recovery processingForward recovery for non-VSAM resources XCF/MRO partner failures Cics recovery processing following a system failure Cics recovery processing following a transaction failure Recovery and restart facilities Cics TS for z/OS 4.1 Recovery and Restart Guide Units of work Resource recovery in CicsShunted units of work Active and retained states for locks Locks Synchronization points Units of work and syncpoints Examples of synchronization points Backout of units of work Cics recovery manager DB2 Coordinating updates to local resources Managing indoubt units of work Coordinating updates in distributed units of work Cics system log Resynchronization after system or connection failureInformation recorded on the system log System activity keypoints Forward recovery logsUser journals and automatic journaling Resource recovery in Cics Cics TS for z/OS 4.1 Recovery and Restart Guide Normal shutdown processing Shutdown and restart recoveryFirst quiesce stage Third quiesce stage Second quiesce stage Shunted units of work at shutdown Warm keypoints Immediate shutdown processing Perform Shutdown Immediate Flushing journal buffersPerform Immediate not recommended Dfhcesd Shutdown requested by the operating system Shutdown assist transaction Uncontrolled termination Global catalog Cataloging Cics resources Local catalog Shutdown initiated by Cics log manager Effect of problems with the system log DFHRM0403 and DFHRM0404 How the state of the Cics region is reconstructed Warm restart Overriding the type of start indicatorEmergency restart About this task Cold start Recovery of data during an emergency restartAn initial start of Cics Dynamic RLS restart Running with persistent sessions support Recovery with Vtam persistent sessionsSNPS, single-node persistent sessions MNPS, multinode persistent sessions Situations in which sessions are not reestablished Situations in which Vtam does not retain sessions Running without persistent sessions support Regions that do have persistent sessions support Cics TS for z/OS 4.1 Recovery and Restart Guide Part 2. Recovery and restart processes Cics TS for z/OS 4.1 Recovery and Restart Guide Starting Cics with the START=COLD parameter Cics cold start Vsam Files Transient data Temporary storageTemporary storage data sharing server Data tables Transactions Resource definitions dynamically installedJournal names and journal models Library resources Monitoring and statistics Committing and cataloging resources installed from the CSDTerminal control resources Distributed transaction resources Installable set installDump table Starting Cics with the START=INITIAL parameter Cics cold start Cics TS for z/OS 4.1 Recovery and Restart Guide Rebuilding the Cics state after a normal shutdown Cics warm restart Data set name blocks Reconnecting to Smsvsam for RLS accessRecreating non-RLS retained locks Trigger levels for Terminal and System only TDINTRA=NOEMPTY the default No autoinstall for programs Start requests Autoinstall for programs Tcam and sequential Bsam devices CSD-defined resource definitions Urimap definitions and virtual hosts Cics TS for z/OS 4.1 Recovery and Restart Guide Recovering information from the system log Recovering after a Cics failureCics emergency restart Driving backout processing for in-flight units of work Other backout processing Effect of delayed recovery on Pltpi processingRebuilding the Cics state after an abnormal termination RLS restart processing and orphan locks Reconnecting to Smsvsam for RLS accessRecreating non-RLS retained locks Start requests Distributed transaction resources Cics TS for z/OS 4.1 Recovery and Restart Guide Cics ARM processing Automatic restart managementRestrictions Waiting for predecessor subsystems Registering with ARMDe-registering from ARM Before you begin ARM couple data sets Failing to registerCics restart JCL and parameters Workload policies Connecting to VtamCics Start options Messages associated with automatic restart Covr transactionAutomatic restart of Cics data-sharing servers Server ARM processing Waiting on events during initialization Server commands for ARM supportServer initialization parameters for ARM support ARMELEMENTNAME=elementname Unit of work recovery Unit of work recovery and abend processingIn-flight-failed Commit-failed Backout-failed Transaction backoutIndoubt-failed Bdam files and Vsam Esds files FilesCics data tables Auxiliary temporary storage Intrapartition transient dataStart requests Start with nonrecoverable data no Protect Start with recoverable data ProtectStart with nonrecoverable data Protect Restart of started transactions Basic mapping support BMS messages Exec Cics Cancel requests Backout-failed recovery Transient data ErrorRetrying backout-failed units of work Disposition of data sets after backout failures Logical delete not performed Open errorSmsvsam server failure Duplicate key error DFSMSdss non-BWO backup in progressSmsvsam server recycle during backout Coupling facility cache structure failure Lock structure full error Commit-failed recoveryNone of the above Indoubt failure recovery Investigating an indoubt failure Cemt Inquire UOW details of UOW AC0CD65E5D990800 Cemt Inquire UOWENQ-details of a lock associated with a UOW Cache failure support Recovery from failures associated with the coupling facility Rebuilding the lock structure Lost locks recoveryNotifying Cics of Smsvsam restart Performing lost locks recovery for failed units of work Connection failure to a coupling facility cache structure MVS system recovery and sysplex recoveryConnection failure to a coupling facility lock structure Exit code Transaction abend processing Transaction restart Abnormal termination of a task Processing operating system abends and program checks Actions taken at transaction failure Unit of work recovery and abend processing Cics TS for z/OS 4.1 Recovery and Restart Guide Terminal error processing Communication error processingNode error program Dfhznep Terminal error program Dfhtep Intersystem communication failures Part 3. Implementing recovery and restart Cics TS for z/OS 4.1 Recovery and Restart Guide Questions relating to recovery requirements Planning aspects of recoveryApplication design considerations Validate the recovery requirements statement End user’s standby procedures Designing the end user’s restart procedureCommunications between application and user System recovery table SRT System definitions for recovery-related functionsResource definitions for recovery Security Transient data queues Documentation and test plansTemporary storage table Program list table PLT Cics TS for z/OS 4.1 Recovery and Restart Guide Defining system and general log streams Defining system log streams Defining log streams to MVSSystem log streams General log streams Without a Journalmodel definition Specifying a Journalmodel resource definitionWith a Journalmodel definition Model log streams for Cics system logs Recovery considerations Sharing system logger structures between 2 MVS images Varying the model log stream name Activity keypointingAbout this task System log scan during restart Keeping system log data to a minimum Moving units of work to the secondary log Log-tail deletion Avoiding retention periods on the system log Writing user-recovery dataRetrieving user records from the system log About this task Procedure Defining forward recovery log streamsLong-running transactions What to do next Model log streams for Cics general logs Defining the log of logs Merging data on shared general log streams Reading log streams offline Log of logs failure Effect of daylight saving time changes Adjusting local timeTime stamping log and journal records Offline utility program, Dfhjup Cics TS for z/OS 4.1 Recovery and Restart Guide Recovery for transactions Defining recoverability for CICS-managed resourcesDefining transaction recovery attributes DTIMOUTNO1-6800 WAITTIME00,00,00dd,hh,mm Indoubt options for distributed transactions File-owning regions and RLS access Recovery for filesVsam files Sharing data sets with batch jobs Defining files as recoverable resources Basic direct access method BdamForward recovery Backward recovery Recoveryall Vsam files accessed in non-RLS mode Inquiring on recovery attributes Vsam files accessed in RLS mode Bdam files File recovery attribute consistency checking non-RLSCSD data set Overriding open failures at the Xfcnrec global user exit Cics responses to file open requests Implementing forward recovery with Cics Vsam Recovery Implementing forward recovery with user-written utilitiesRecovery for intrapartition transient data Backward recovery Physical recovery Logical recovery No recovery Forward recovery Input extrapartition data sets Recovery for extrapartition transient data Recovery for temporary storage Using post-initialization Pltpi programsOutput extrapartition data sets Configuring Cics to support persistent messages Recovery for Web services Results What to do next Defining local queues in a service provider Persistent message processing Error processing Defining recoverability for CICS-managed resources Cics TS for z/OS 4.1 Recovery and Restart Guide Designing applications for recovery Programming for recoverySplitting the application into transactions Cics TS for z/OS 4.1 Recovery and Restart Guide Program design SAA-compatible applicationsDividing transactions into units of work Conversational processing Processing dialogs with usersPseudoconversational processing Mechanisms for passing data between transactions Cics recoverable resourcesMain storage areas Designing to avoid transaction deadlocks Implications of interval control Start requests Implications of automatic task initiation TD trigger level Using transient data queuesTerminal paging through BMS Transaction failures Managing transaction and system failures Exec Cics Syncpoint Rollback command Handle Abend commandsDynamic transaction backout System failures Use of the program error program DfhpepHandling abends and program level abend exits Transaction restart after DTB Processing the Ioerr condition Command Information provided PL/I programs and error handling Start Transid commandsLocking enqueuing on resources in application programs Implicit locking for files Nonrecoverable files Recoverable files Cics TS for z/OS 4.1 Recovery and Restart Guide Implicit enqueuing on recoverable temporary storage queues Implicit enqueuing on logically recoverable TD destinationsWriteq TD Readq TD Deleteq TD Explicit enqueuing by the application programmer Implicit enqueuing on DL/I databases with DbctlDirect methods HDAM, Hidam Sequential methods HSAM, HISAM, Shisam Deadlock Possibility of transaction deadlock User exits for transaction backout Where you can add your own code Xrcinpt exit Xrcinit exitXfcbfail global user exit Xfcbover global user exit Xfcldel global user exitXfcbout global user exit Coding transaction backout exits CICS-supplied PEP Using a program error program PEP Your own PEP Omitting the PEP Cics TS for z/OS 4.1 Recovery and Restart Guide Quiescing RLS data sets Resolving retained locks on recoverable resources Illustration of the quiesce flow across two Cics regions RLS quiesce and unquiesce functions MVS1 MVS2 Cics AOR1 Cics AOR2 Other quiesce interface functions Non-BWO data set backup start BWO backup start Non-BWO data set backup endBWO backup end Forward recovery complete Lost locks recovery complete Switching from RLS to non-RLS access modeException for read-only operations Quiesce coupling facility cache available What can prevent a switch to non-RLS access mode? Investigating which retained locks are held and why Inquire Dsname Resolving retained locks and preserving data integrity Shcds List subcommands Choosing data availability over data integrity Batch-enabling sample programs Cemt command examples Uowdsnfail Denynonrlsupdate subcommand Permitnonrlsupdate subcommandSpecial case lost locks Overriding retained locks Post-batch processing Coupling facility data table retained locks Procedure for moving a data set with retained locks Moving recoverable data sets that have retained locksUsing the Repro method Shcds Frsetrr Using the Export and Import functions Rebuilding alternate indexes Forward recovery of data sets accessed in RLS mode Forward recovery procedures Recovery of data set with volume still available Issue the Frbind subcommand Recovery of data set with loss of volumeIssue the Frresetrr subcommand Alter the new data set name Volume recovery procedure using Cfvol Quiesce Forward recovery procedures Example of recovery using data set backup Shcds LISTSUBSYSDS’ADSWA01D’ Vary SMS,SMSVSAM,FORCEDELETELOCKSTRUCTURE Inquire Uowdsnfail Status Results Example of recovery using volume backup Above command produced the following message Catalog recovery Procedure for failed RLS mode forward recovery operation Forward recovery of data sets accessed in non-RLS modeClose all files Forward recovery procedures Uowactioncommit Backout Force Procedure for failed non-RLS mode forward recovery operation Cics TS for z/OS 4.1 Recovery and Restart Guide BWO and backups Backup-while-open BWOBWO and concurrent copy Component name Full DFSMS/MVS name Previous product Name BWO requirements Which data sets are eligible for BWO Hardware requirementsVsam control interval or control area split How you request BWO Specifying BWO using access method servicesResults Specifying BWO on Cics file resource definitions Removing BWO attributes Systems administrationBatch jobs BWO processing File opening First file opened in non-RLS mode against a clusterBack-level data sets Subsequent files opened when use count is zero Subsequent files opened when use count is not zero Restriction for Vsam upgrade set File closing non-RLS mode Shutdown and restart Data set backup and restoreControlled shutdown Immediate or uncontrolled shutdown Invalid state changes for BWO attributes Vsam access method services Data set restore Forward recovery loggingData sets Non-SMS managed storage Recovery point non-RLS mode Recovering Vsam spheres with AIXs ASM XOPTSCICS,NOEPILOG,SP An assembler program that calls Dfsms callable services DS 8C MVC MVC SUCMSG18,SUCTXT1 SUCMSG21,SUCTXT2 Cics TS for z/OS 4.1 Recovery and Restart Guide Why have a disaster recovery plan? Disaster recovery Disaster recovery testing Tier 0 no off-site data Six tiers of solutions for off-site recoveryTier 1 physical removal Disaster recovery tier 1 physical removal Tier 3 electronic vaulting Tier 2 physical removal with hot siteTier Summarizes the tier 3 solution Tier 0-3 solutions Tier 0 Tier 4 active secondary site Disaster recovery tier 4 active secondary site Tier 6 minimal to zero data loss Tier 5 two-site, two-phase commit Summarizes the tier 6 solution Tier 4 Tier 4-6 solutions Peer-to-peer remote copy Pprc and extended remote copy XRC Disaster recovery and high availability Use Pprc for high value transactions Pprc or XRC? Other benefits of Pprc and XRC Remote Recovery Data Facility Disaster recovery personnel considerations Choosing between Rrdf and 3990-6 solutions MVS system logger recovery support Disaster recovery facilitiesReturning to your primary site Remote Recovery Data Facility support Cics Vsam Recovery Qsam copyRemote site recovery for RLS-mode data sets Cics VR shadowing Final summary 241 Part 4. Appendixes Cics TS for z/OS 4.1 Recovery and Restart Guide 243 Trademarks Cics books for Cics Transaction Server for z/OS Bibliography CICSPlex SM books for Cics Transaction Server for z/OS Administration and ManagementOther Cics publications Accessibility Cics TS for z/OS 4.1 Recovery and Restart Guide Index Airdelay DL/I 251 Tpend Readers’ Comments Wed Like to Hear from You Readers’ Comments Wed Like to Hear from You Page SC34-7012-01