IMS Disaster Recovery



Similar documents
IMS Disaster Recovery Overview

IMS Disaster Recovery Overview

Once the RTO and RPO objectives are known, the customer is able to determine which disaster recovery methodology works best for their needs.

IT Service Management

Simplify and Improve Database Administration by Leveraging Your Storage System. Ron Haupert Rocket Software

Storage System High-Availability & Disaster Recovery Overview [638]

Lisa Gundy IBM Corporation. Wednesday, March 12, 2014: 11:00 AM 12:00 PM Session 15077

Introduction to Enterprise Data Recovery. Rick Weaver Product Manager Recovery & Storage Management BMC Software

Informix Dynamic Server May Availability Solutions with Informix Dynamic Server 11

Exploiting the Virtual Tape System for Enhanced Vaulting & Disaster Recovery A White Paper by Rocket Software. Version 2.

FICON Extended Distance Solution (FEDS)

IBM TotalStorage IBM TotalStorage Virtual Tape Server

HITACHI DATA SYSTEMS USER GORUP CONFERENCE 2013 MAINFRAME / ZOS WALTER AMSLER, SENIOR DIRECTOR JANUARY 23, 2013

DB2 for z/os System Level Backup Update

Recommendations from DB2 Health Check Studies: Speed of Recovery

Alternate Methods of TSM Disaster Recovery: Exploiting Export/Import Functionality

Big Data Storage in the Cloud

BMC Mainframe Solutions. Optimize the performance, availability and cost of complex z/os environments

IBM DB2 Recovery Expert June 11, 2015

Application Backup and Restore using Fast Replication Services. Ron Ratcliffe March 13, 2012 Session Number 10973

User Experience: BCPii, FlashCopy and Business Continuity

IBM Systems and Technology Group Technical Conference

DASD Backup Automation

Virtual Tape Systems for IBM Mainframes A comparative analysis

Secure Database Backups with SecureZIP

HA / DR Jargon Buster High Availability / Disaster Recovery

z/os Data Replication as a Driver for Business Continuity

Backup Solutions with Linux on System z and z/vm. Wilhelm Mild IT Architect IBM Germany mildw@de.ibm.com 2011 IBM Corporation

BACKUP SECURITY GUIDELINE

Contents. SnapComms Data Protection Recommendations

Business Resilience for the On Demand World Yvette Ray Practice Executive Business Continuity and Resiliency Services

IBM and VERITAS Practical disaster recovery

CA High Performance Recovery for IMS for z/os

IP Storage On-The-Road Seminar Series

Tributary Systems Storage Director Provides Superior ROI. By Ed Ahl, Director of Business Development

IBM Virtualization Engine TS7700 GRID Solutions for Business Continuity

DB2 for z/os Backup and Recovery: Basics, Best Practices, and What's New

Disaster Recovery for Oracle Database

Remote Copy Technology of ETERNUS6000 and ETERNUS3000 Disk Arrays

Best Practices for DB2 on z/os Backup and Recovery

Data Protection History, Evolution, Best Practices By Global Data Vault

Availability and Disaster Recovery: Basic Principles

High Availability and Disaster Recovery for Exchange Servers Through a Mailbox Replication Approach

Affordable Remote Data Replication

Disaster Recovery Feature of Symfoware DBMS

Synchronous Data Replication

Volvo IT s Mainframe journey

Four Steps to Disaster Recovery and Business Continuity using iscsi

Data Protection with IBM TotalStorage NAS and NSI Double- Take Data Replication Software

How To Improve Your Database Performance

Protecting Microsoft SQL Server

Solutions for BC/DR. (as seen trough Storage) IBM System Storage TM. Klemen Bačak , Banja Luka

Disaster Recovery Solutions for Oracle Database Standard Edition RAC. A Dbvisit White Paper

VERITAS Business Solutions. for DB2

DB2 for z/os: [Some] Backup and Recovery Utility Enhancements in V8/9/10

We look beyond IT. Cloud Offerings

DISASTER RECOVERY BUSINESS CONTINUITY DISASTER AVOIDANCE STRATEGIES

Implementing Disaster Recovery? At What Cost?

Using the NDMP File Service for DMA- Driven Replication for Disaster Recovery. Hugo Patterson

How To Choose A Business Continuity Solution

High Availability and Disaster Recovery Solutions for Perforce

Storage Virtualization for Mainframe DASD, Virtual Tape & Open Systems Disk

REMOTE BACKUP-WHY SO VITAL?

Backup with synchronization/ replication

BROCADE PERFORMANCE MANAGEMENT SOLUTIONS

Data Center Optimization. Disaster Recovery

High Availability with Postgres Plus Advanced Server. An EnterpriseDB White Paper

What You Need to Know About Cloud Backup: Your Guide to Cost, Security, and Flexibility

Deduplication and Beyond: Optimizing Performance for Backup and Recovery

SAN Conceptual and Design Basics

System Infrastructure Non-Functional Requirements Related Item List

The Economic Impact Of Zero Data Loss. White Paper

Why cloud backup? Top 10 reasons

A SURVEY OF POPULAR CLUSTERING TECHNOLOGIES

Objectif. Participant. Prérequis. Pédagogie. Oracle Database 11g - Administration Workshop II - LVC. 5 Jours [35 Heures]

Program Product. Open System Backup

Contingency Planning and Disaster Recovery

Continuous Data Protection for DB2

IBM TSM DISASTER RECOVERY BEST PRACTICES WITH EMC DATA DOMAIN DEDUPLICATION STORAGE

Introduction. Ease-of-Use

An Introduction to z/os DASD Backup and Archiving

SHARE Lunch & Learn #15372

Backup and Recovery 1

WHAT IS DISASTER RECOVERY

Backup Exec Private Cloud Services. Planning and Deployment Guide

Frequently Asked Questions (FAQs) United Computer Group, Inc. VAULT400 System i (AS/400) Agent

Virtual Infrastructure Security

WHITE PAPER: ENTERPRISE SOLUTIONS. Veritas Volume Replicator Option by Symantec A Guide to Understanding Volume Replicator

Transcription:

IMS Disaster Recovery Part 1 Understanding the Issues February 5, 2008

Author Bill Keene has almost four decades of IMS experience and is recognized world wide as an expert in IMS recovery and availability. Prior to retiring from IBM, Bill was the technical leader of the IMS unit within the Dallas Systems Center, part of IBM s Advanced Technical Support organization. In that capacity, Bill presented at numerous IBM and user technical conferences and was the technical focal point within IBM s marketing division for product introduction programs dealing with XRF, RRDF, and RSR, all IMS functions dealing with local system or disaster recovery. Bill retired from IBM in 1998 and is currently utilizing his IMS database recovery knowledge and expertise as the product architect of irecover and ichange, the primary components in the IMS database recovery solutions provided by NEON Enterprise 2008 NEON Enterprise Software, Inc. 1

Agenda Disaster Recovery Basics Hardware Solutions GDPS (Geographically Dispersed Parallel Sysplex) DASD remote copy (Metro Mirror / z/os Global Mirror) Remote tape Software Solutions IMS RSR E-Net RRDF Manual PTAM (Pickup Truck Access Method) Database recovery considerations 2008 NEON Enterprise Software, Inc. 2

Disaster Recovery Basics Objective: Restore IMS service at a remote location when the local (normal) computing facility is no longer available Remoteness is in the eye of the beholder In general, the further away, the better Ensure the survival of the business! 2008 NEON Enterprise Software, Inc. 3

Disaster Recovery Basics Disaster Recovery Methodology Determined by business needs How remote is remote (type of disaster) How much data can business lose? How long can business go without service? What can business afford? Is DR methodology allowed to impact local availability? Cost goes up as data loss and service unavailability go down Not linear 2008 NEON Enterprise Software, Inc. 4

GDPS Availability solution based on automation and Metro Mirror / z/os Global Mirror Timer Timer Sys A Sys B 100 km Sys C Sys D Remote Copy 2008 NEON Enterprise Software, Inc. 5

GDPS Advantages Transparent to application software (IMS, DB2, VSAM) No data loss (Metro Mirror) Assumes no connectivity or write failures No impact on local operations Assumes no connectivity or write failures No database recovery operations required Assumes no connectivity or write failures Consistent recovery procedures Quick workload transfer No subsystem coordination problems 2008 NEON Enterprise Software, Inc. 6

GDPS Disadvantages Limited distance (100 km 40 km between Sysplex Timers) Duplicate DASD RECONs, OLDS, WADS, and databases must participate Impacts IMS throughput and response times PPRC elongates DASD write response times (10 microsec/km) High bandwidth requirements Backup site must always be available Logs archived to tape are a challenge IMS 3494 Magstar VTS duplexing function might help Other techniques still desirable for backup 2008 NEON Enterprise Software, Inc. 7

DASD Synchronous Remote Copy (Metro Mirror) DASD subsystem function Application Subsystem 1. Write to primary volume 2. Disconnect from channel (CE) 3. Write to secondary volume 4. Signal write complete 5. Post I/O completion (DE) 1. 2. 5. Local Site Storage Control 3. 4. Remote Site Storage Control 2008 NEON Enterprise Software, Inc. 8

DASD Asynchronous Remote Copy (z/os Global Mirror) DASD subsystem / DFSMS function System Data Mover Local Site Storage Subsystem Remote Site Storage Subsystem 2008 NEON Enterprise Software, Inc. 9

DASD Remote Copy What to copy? Databases? No data loss (Metro Mirror) / minimal loss (z/os Global Mirror) RECONs, OLDS, and WADS must also be copied No database recoveries needed In-flight updates must be backed out SLDS / RLDS and image copies assumed to be at remote site OLDS / WADS? No data loss (Metro Mirror) / minimal data loss (z/os Global Mirror) RECONs must also be copied Database recoveries needed SLDS / RLDS and image copies assumed to be at remote site 2008 NEON Enterprise Software, Inc. 10

DASD Remote Copy What to copy? SLDS / RLDS Updates after last SLDS / RLDS lost RECONs must also be copied Database recoveries needed Image copy data assumed to be at remote site Questionable RECONs? SLDS / RLDS and image copy data assumed to be at remote site Updates after last SLDS / RLDS at remote site lost Database recoveries needed Doesn t make sense 2008 NEON Enterprise Software, Inc. 11

DASD Remote Copy Advantages Transparent to application software (IMS, DB2, VSAM) Depends on what is copied No (Metro Mirror) / minimal (z/os Global Mirror) data loss if OLDS / WADS copied Assumes no connectivity or write failures No impact on local operations Assumes no connectivity or write failures No database recovery operations required Assumes databases copied Assumes no connectivity or write failures 2008 NEON Enterprise Software, Inc. 12

DASD Remote Copy Advantages Consistent recovery procedures Depends on what is copied Quick workload transfer Depends on what is copied No subsystem coordination problems Depends on what is copied 2008 NEON Enterprise Software, Inc. 13

DASD Remote Copy Disadvantages Metro Mirror has distance limitations (300 km) Duplicate DASD Depends on what is copied Metro Mirror impacts IMS throughput and response times DASD write response time elongated (10 microsec / km) Depends on what is copied High bandwidth requirements Depends on what is copied Image copies might be required following takeover Depends on log and image copy availability 2008 NEON Enterprise Software, Inc. 14

DASD Remote Copy Disadvantages Logs archived to tape are a challenge IBM TotalStorage 3494 Magstar VTS duplexing function might help Other techniques still desirable for backup Not as simple as it sounds IBM System Storage DS8000: Copy Services with IBM System z (800+ pages) IBM TotalStorage Peer-to-Peer Virtual Tape Server: Planning and Implementation Guide (450+ pages) 2008 NEON Enterprise Software, Inc. 15

Remote Tape What is it? Tape subsystem at remote location connected to local system via channel extender technology Appears to be a local device Used for Secondary image copy data sets Secondary archived log data sets RECON backup data sets Can use NEON irecover or NEON ichange to create preconditioned Load and Go RECON backup data sets 2008 NEON Enterprise Software, Inc. 16

Remote Tape Advantages Low tech / low cost Easy to implement and maintain Image copy and archived log data off-site when created RECON backups off-site when created No impact on IMS throughput and response times No duplicate DASD Subsystem coordination possible with the right software NEON irecover and NEON ichange Can be used with other techniques 2008 NEON Enterprise Software, Inc. 17

Remote Tape Disadvantages Medium data loss Data since last archived log unavailable Database recovery operations required Image copy times could be elongated Log archive times could be elongated Procedures needed to catalog remote tape files at remote location 2008 NEON Enterprise Software, Inc. 18

IMS Remote Site Recovery (RSR) IMS feature Recovery log tracking (RLT) Database log tracking (DLT) Electronic log vaulting (DLT and RLT) Log blocks transmitted to special purpose IMS system at remote location prior to writing to OLDS (on-line) / RLDS (batch) IMS log records contain RECON update information Shadow databases (DLT) Option to maintain shadow databases Continuous forward recovery Selectable by database 2008 NEON Enterprise Software, Inc. 19

IMS RSR Advantages Minimal data loss Updates since last OLDS write Updates in transmission pipeline Minimal bandwidth No impact on IMS throughput and response times No distance limitations No database recovery operations required when shadow databases are maintained Some subsystem coordination support provided 2008 NEON Enterprise Software, Inc. 20

IMS RSR Disadvantages Only supports IMS data IMS system (CPU cycles) required at remote location Requires about 5% of active site CPU usage for RLT Requires about 10 15% of active site CPU usage for DLT Does not address image copy data set transport 2008 NEON Enterprise Software, Inc. 21

E-Net RRDF Electronic log vaulting Log blocks transmitted to RRDF Receive system at remote location after written to OLDS (on-line) / RLDS (batch) RECON updates transmitted to RRDF Receive system at remote location Works quite differently from IMS RSR! 2008 NEON Enterprise Software, Inc. 22

E-Net RRDF Advantages Minimal data loss Updates since last OLDS write Can force OLDS write during low logging periods Updates in transmission pipeline Minimal bandwidth No distance limitations Support for IMS, DB2, and CICS Common electronic log vaulting solution 2008 NEON Enterprise Software, Inc. 23

E-Net RRDF Disadvantages RRDF Receive system required at remote location IMS database recovery operations required Does not address image copy data set transport 2008 NEON Enterprise Software, Inc. 24

PTAM (Pickup Truck Access Method) What is it? Copies of recovery media periodically shipped off-site Secondary image copies Secondary logs Change accums? RECON backups Can use NEON irecover or NEON ichange to create preconditioned Load and Go RECON backup data sets Frequency of shipments depends on amount of acceptable data loss Data not really available until the truck arrives at the remote location! 2008 NEON Enterprise Software, Inc. 25

PTAM Advantages Low tech / low cost Easy to implement and maintain Minimal resources at remote location No impact on IMS throughput and response times No duplicate DASD Subsystem coordination possible with the right software NEON irecover and NEON ichange Can be used with other techniques 2008 NEON Enterprise Software, Inc. 26

PTAM Disadvantages Maximum data loss Data not hardened until truck leaves local site Data not available until truck arrives at remote site Error prone Manual procedures to collect data at local site Manual procedures to receive data at remote site Need software that can create a pick list of required data sets NEON irecover NEON ichange 2008 NEON Enterprise Software, Inc. 27

Disaster Recovery Requirements Local system data availability should not be impacted by disaster recovery methodology RECON conditioning / cleanup operations should be automatic Database recovery speed One pass, parallel recovery operations are a necessity Image copy, index rebuild, and hashed pointer checking operations need to be performed as part of the recovery process User needs complete control over the time to which databases are recovered Coordinated recovery with DB2 is mandatory 2008 NEON Enterprise Software, Inc. 28

Recovery Considerations IMS and DB2 log data available at remote site Any time stamp is possible with NEON irecover Process Restore conditioned RECON backup created by NEON irecover Recover IMS databases to ANY desired time stamp using NEON irecover Single step operation IC, PC, and index rebuild performed while recovering Recoveries performed with transactional consistency Perform DB2 conditional restart to same time stamp (DB2 facility) Use in-doubt DB2 UOR status report from NEON irecover to resolve DB2 in-doubt UORs from IMS UORs 2008 NEON Enterprise Software, Inc. 29

NEON irecover DR Support Steps to accomplish disaster recovery Restore latest DR RECON backup //DR1 JOB... //STEP1 EXEC PGM=NSRRCNRS //STEPLIB DD DSN=... //BACKUP DD DSN=... //RECON1 DD DSN=... Restore RECON duality //DR2 JOB... //STEP1 EXEC PGM=DSPURX00 //STEPLIB DD DSN=... //SYSPRINT DD SYSOUT=* LIST.RECON STATUS /* 2008 NEON Enterprise Software, Inc. 30

NEON irecover DR Support Steps to accomplish disaster recovery Run Eclipse irecover to desired time stamp //DR3 JOB... //STEP1 EXEC PGM=NSRMAIN //STEPLIB DD DSN=... //NEONIN DD * RECOVERGRP DBD(*) RECOVERTO(...) /* Go get coffee! 2008 NEON Enterprise Software, Inc. 31

Alternatives to Log Data Environment User needs only a few DR points per day Recovering to prior 2 AM is good enough User does not want to quiesce database update activity or data sharing level when establishing a DR point User wishes to send only image copy and change accum data sets off-site (no logs), along with a RECON backup User wishes to recover databases to a consistent state (prior 2 AM) at the disaster site Quickly, correctly, with minimum effort 2008 NEON Enterprise Software, Inc. 32

NEON ichange DR Support Special DR change accumulation data sets Normal change accumulation information Additional records, if needed, to restore consistency Load and Go RECON backup Conditioned for disaster recovery All necessary cleanup operations already performed DR media inventory report Lists data sets required for disaster recovery Can be used to ensure all required assets are off-site DR coverage report Lists all recoverable data sets that are not in a DR CA 2008 NEON Enterprise Software, Inc. 33

NEON ichange DR Support Steps to establish a 2:00 AM DR point at the local site Automation issues /SWI OLDS immediately after 2:00 AM Automation submits Eclipse ichange job after archives complete //DRCA JOB... //STEP1 EXEC PGM=NSAMAIN //STEPLIB DD DSN=... //NEONIN DD * ACCUMGRP GROUP(CAGRP1) ACCUMTO(* 02:00:00) ACCUMGRP GROUP(CAGRP2) ACCUMTO(* 02:00:00) /* DR change accums and RECON backup sent off-site, along with any image copy data sets created since last DR point 2008 NEON Enterprise Software, Inc. 34

Eclipse ichange Steps to accomplish disaster recovery Restore latest DR RECON backup //DRCA JOB... //STEP1 EXEC PGM=NSARCNRS //STEPLIB DD DSN=... //BACKUP DD DSN=... //RECON1 DD DSN=... Restore RECON duality //DR2 JOB... //STEP1 EXEC PGM=DSPURX00 //STEPLIB DD DSN=... //SYSPRINT DD SYSOUT=* //SYSIN DD * LIST.RECON STATUS /* 2008 NEON Enterprise Software, Inc. 35

Eclipse ichange Steps to accomplish disaster recovery Run Eclipse irecover //DRCA JOB... //STEP1 EXEC PGM=NSRMAIN //STEPLIB DD DSN=... //NEONIN DD * RECOVERGRP GROUP(CAGRP1) RECOVERGRP GROUP(CAGRP2) /* Go get coffee! 2008 NEON Enterprise Software, Inc. 36

Summary DR solution is a business decision How much data loss is acceptable How long a data outage is acceptable How much expense is acceptable DR solution needs to accommodate both IMS and DB2 All solutions require recovery software that can Recover thousands of objects in a single step and survive Recover objects to any point in time Support multiple recovery points in time if needed Recovery solution needs to be designed with disaster recovery in mind 2008 NEON Enterprise Software, Inc. 37

Questions and Answer Period

Reminder! To find out more about irecover and ichange and all the magical things they can do, tune in next meeting (April 8 th ) 2008 NEON Enterprise Software, Inc. 39