book

IBM Db2 Analytics Accelerator V7 High Availability and Disaster Recovery

by Ute Baumbach, Frank Neumann

October 2020

Intermediate to advanced

54 pages

1h 51m

English

IBM Redbooks

Read now

Unlock full access

Trademarks
AuthorsNow you can become a published author, too!Comments welcomeStay connected to IBM Redbooks
1.1 Recovery time objective1.2 Recovery point objective
2.1 Storage failures and synchronous and asynchronous data mirroring2.2 Db2 Active-Active configuration with GDPS Metro2.3 Db2 Active-Passive configuration with GDPS Global2.4 Db2 Active-Passive configuration with GDPS Metro2.5 Db2 multi-site configurations and tests
3.1 Workload balancing3.1.1 Directing queries to specific accelerators3.2 Data maintenance and synchronization with multiple accelerators3.2.1 Loading into multiple accelerators by using the ACCEL_LOAD_TABLES stored procedure3.2.2 Loading into multiple accelerators using Db2 Analytics Accelerator Loader3.2.3 Loading into multiple accelerators using the HA Load sample program of Db2 Analytics Accelerator3.2.4 Loading a subset of data3.2.5 Using accelerator-only tables with multiple accelerators3.2.6 Loading accelerator only data with Db2 Analytics Accelerator Loader3.2.7 Incremental Update into multiple accelerators3.3 HA setup for incremental update3.3.1 HA setup for incremental update using Integrated Synchronization3.3.2 HA setup for incremental update using CDC3.4 High-performance storage saver and multiple accelerators3.5 Federation with multiple accelerators3.6 Definition of accelerators in Db2 catalog tables
4.1 Combining V5 and V7 Accelerators4.2 Similarities and differences of IIAS and Z deployments for V7 Accelerators
5.1 Stand-by accelerators5.2 Network setup options5.3 Latency and bandwidth requirements5.4 Db2 active-active configuration with GDPS Metro and active accelerators5.4.1 Data maintenance by using regular load cycles5.4.2 Data maintenance by using incremental update5.4.3 High-performance storage saver5.5 Db2 active-standby configuration with GDPS Metro and active accelerators5.6 Db2 active-standby configuration with GDPS Global and active accelerators5.6.1 Data maintenance by using regular load cycles5.6.2 Data maintenance by using incremental update5.6.3 High-performance storage saver5.7 Db2 active-standby configuration with GDPS Global and limited bandwidth to remote accelerator5.8 Combined Db2 active-active-standby environment with GDPS Metro and GDPS Global and accelerators5.9 Built-in HA features within IBM Db2 Analytics Accelerator on IIAS
6.1 IBM Z integration, network, and storage management6.2 Network setup options6.3 Latency and bandwidth requirements6.4 Storage failures and HyperSwap6.5 Active-passive configuration with standby Accelerator and GDPS Metro6.6 Active-active configuration with capacity growth6.7 Multi-site configurations with active and passive accelerators and GDPS

Overview

IBM® Db2® Analytics Accelerator is a workload optimized appliance add-on to IBM DB2® for IBM z/OS® that enables the integration of analytic insights into operational processes to drive business critical analytics and exceptional business value. Together, the Db2 Analytics Accelerator and DB2 for z/OS form an integrated hybrid environment that can run transaction processing, complex analytical, and reporting workloads concurrently and efficiently.

With IBM DB2 Analytics Accelerator for z/OS V7, the following flexible deployment options are introduced:

Accelerator on IBM Integrated Analytics System (IIAS): Deployment on pre-configured hardware and software
Accelerator on IBM Z®: Deployment within an IBM Secure Service Container LPAR

For using the accelerator for business-critical environments, the need arose to integrate the accelerator into High Availability (HA) architectures and Disaster Recovery (DR) processes. This IBM Redpaper™ publication focuses on different integration aspects of both deployment options of the IBM Db2 Analytics Accelerator into HA and DR environments. It also shares best practices to provide wanted Recovery Time Objectives (RTO) and Recovery Point Objectives (RPO).

HA systems often are a requirement in business-critical environments and can be implemented by redundant, independent components. A failure of one of these components is detected automatically and their tasks are taken over by another component. Depending on business requirements, a system can be implemented in a way that users do not notice outages (continuous availability), or in a major disaster, users notice an outage and systems resume services after a defined period, potentially with loss of data from previous work.

IBM Z was strong for decades regarding HA and DR. By design, storage and operating systems are implemented in a way to support enhanced availability requirements. IBM Parallel Sysplex® and IBM Globally Dispersed Parallel Sysplex (IBM GDPS®) offer a unique architecture to support various degrees of automated failover and availability concepts.

This IBM Redpaper publication shows how IBM Db2 Analytics Accelerator V7 can easily integrate into or complement existing IBM Z topologies for HA and DR.
If you are using IBM Db2 Analytics Accelerator V5.1 or lower, see IBM Db2 Analytics Accelerator: High Availability and Disaster Recovery, REDP-5104.