+ All Categories
Home > Documents > Presenter’s Name Presenter’s Title Smoothing the DB2 version migration path!

Presenter’s Name Presenter’s Title Smoothing the DB2 version migration path!

Date post: 16-Dec-2015
Category:
Upload: reese-reason
View: 228 times
Download: 0 times
Share this document with a friend
Popular Tags:
25
Presenter’s Name Presenter’s Title Smoothing the DB2 version migration path!
Transcript

Presenter’s Name Presenter’s Title

Smoothing the DB2 version migration path!

Migrating to DB2 v10?

Do you want to avoid surprises and delays during and after the upgrade

process?

DB2 subsystem clones offer multiple levels of testing environments

These clones can be created with minimum skilled

resources, minimum time and no interruption to your production systems.

Creating exact copies of production quickly for high quality testing reduces the risk of impacting critical

applications during the migration process.

2

AgendaWhat is lurking in production? (and other environments)

Problems and concerns

Test environments needed for successSmoothing the migration path

Test environments created by VCRDatabase storage integration

Cloning overview

Cloning terminologyClone DB2 systems

Step by Step test environment creation

3

4

You are migrating to a new version of

DB2!

4

What’s lurking in production?

5

But, what is lurking in your production

environment?

5

What’s lurking in production?

6

What is waiting to be uncovered by the

new software?

6

What’s lurking in production?

7

What’s lurking in production?

Your environment is unique – How do those new features really work in YOUR production environment?

How do those WAD elements mesh with YOUR production environment?

Are there bugs that can only surface in YOUR production environment?

Are there issues that can only manifest with the combinations of data and data volumes in YOUR production environment?

What are the unexpected events for which you cannot test because they cannot be ANTICIPATED with your current knowledge?

Is there legacy code using undocumented features that no longer exist? Is there Assembler code using addresses that are now used by DB2? Is there a clever SQL technique that no longer works? Is there code that was never invoked because the data volumes never

reached the levels and throughput possible with the new DB2? Something else?

7

Development environments may be even more complicated

8

What’s lurking in development?

Development environmentsCritical applications may be in test – with new

features – DB2 VNEXT may impactPerformance testing can be impacted by new

versions of DB2 Optimizer changes, etc

Multiple copies of the same application objects and data – Make more complicated environment to manage

Multiple versions of the same applicationNew features of an application that have not

quite made it into production

8

9

How do you find out before it bites you?

Upgrading to a new DB2 version is a high impact effort. Any issue could be visible system wide or company wide

Sandbox environments are used to initially test DB2 version release. However, software is rarely kept in sync with production. This is due to resources issues of time, money, space, manpower,

and availability of required skills. Usually “bare bones” and not reflecting production

There is only one Sandbox but many production DB2 configurations.

It is difficult to measure the impact or savings of implementing new DB2 features without testing in a production-like environment.

9

What’s lurking in production?

10

A clone of your production DB2 can

uncover what’s hidden!

10

What’s lurking in production?

11

Smoothing the migration path

Cloning DB2 is not a new concept

There are 2 basic traditional methods The first method – Resource intensive:

Creates a new DB2 system (the clone) Creates the objects from a source DB2 on the new DB2 via DDL Unloads the tables of the source DB2 Loads the table rows of the source DB2 onto the tables of the clone

The second method – Requires an isolated LPAR: Dump the data sets of a source DB2 to tape or some movable media On an isolated LPAR restore the data sets of the source DB2 Start the clone DB2

A third and non-traditional method uses cloning automation and makes the process - Safer Faster Non resource intensive Can clone on the same or on a shared LPAR Uses storage technology which you already own

11

12

Smoothing the migration path

Migration with a DB2 subsystem clone

Create the DB2 subsystem Clone. There is cloning automation which can do this in under an hour

Migrate the clone to the new version of DB2

Execute a parallel set of production jobs on the Clone

If any issues are uncovered, correct them at leisure during business hours, with no fear of an emergency back off at some early morning hour

Apply any necessary patches and migrate the production DB2 with confidence!

12

13

How does it clone so fast?Database and Storage Integration

13

MainframeDatabase Systems

Storage AwareDatabase Tools

Application and Database Management

Domain

Storage Administration and

Business ContinuityDomain

• Organizational Integration• New Backup Methods• New Recovery Strategies• Business Recovery Monitoring• Cloning Automation• Disaster Restart Solutions

SourceDatabase Backup,

Clone,DR

14

Cloning Terminology

A clone is an exact but independent replica Clone a DB2 system by volume

Clone a table space by data set

DB2 system cloning and table space refresh The act of replicating the data, making the replica accessible, and then using the replica in lieu of the

original data

DB2 system cloning automation Clones a complete DB2 system including all its databases

Lowest level is by storage volume

DB2 table and index space refresh automation Refreshes specific table and index spaces

Lowest level is by data set

14

15

Types of DB2 clones

Full Subsystem Clone 2 sources:

DB2 Subsystem storage volumes System Level Backup (SLB)

Partial subsystem Clone Selected storage volumes (partial application data)

Skeleton Clone DB2 directory, catalog, BSDS, and active logs storage

volumes without application data Populated by dataset

Table Space Clones By dataset/object definition

15

16

Fast Replication Data Copy Options

Volume based fast replication options for DB2 system cloning FlashCopy (IBM,EMC,HDS)

SnapShot (IBM,STK)

TimeFinder/Clone Volume Snap (EMC)

TimeFinder/Snap (EMC)

Mirror processesPPRC (IBM,EMC,HDS)

TimeFinder/Mirror, SRDF (EMC)

ShadowImage HUR (HDS)

Data set based fast replication options for DB2 table space refresh Data Set FlashCopy (IBM,EMC,HDS)

Data set SnapShot (IBM,STK)

TimeFinder/Clone Data set Snap (EMC)

16

Fast copy processes offloaded to the storage processorNo host CPU or I/O resources

Fast ReplicationCommands from

z/OS

17

Host Based Data Copy Options

Volume copy options for DB2 system cloning TDMF (IBM)

FDRPAS (Innovation Data Processing)

DFSMSdss (IBM)

FDR (Innovation Data Processing)

Data set copy options for DB2 table space refresh Any traditional data set copy processes

17

Data copy processes use host based CPU and I/O facilitiesSlower than storage-based fast replication

Host-basedCopy Process

18

DB2 Full Subsystem Cloning Steps

18

Correct DB2 catalog and directory page spaces

Production DB2‘Source’

DB2Clone

Target DB2

6

11

Start DB2 in maintenance mode for metadata management

12 Start DB2 clone in normal mode

1 DB2 volumeSelection or SLB

2

Volume copy3

A. SET LOG LOAD(0) SET LOG SUSPEND (*not for an SLB source)

B. Consistency Group

4 SET LOG RESUME if 2A

7

8

9

10

Stop target in maintenance mode

Correct application page spaces

Update DB2 directory and BSDSs

Update DB2 catalog

Rename 5

DB2

SourceDatabaseVolumes

CloneDatabaseVolumes

19

Partial Subsystem Clone

DB2 subsystem is cloned + specific applications

Some application datasets are excluded from renames – and optionally deleted

SourceDB2 Subsystem

Volumes

CloneDB2 Subsystem

Volumes

Application Volumes

Application Volumes

Application Volumes

Application Volumes

20

Populated Subsystem Skeleton Clone

DB2 subsystem is cloned + selected DB2 data sets

Volumes containing application data are NOT copied

SourceDB2 Subsystem

Volumes

CloneDB2 Subsystem

Volumes

Application Volumes

Application Volumes

Application Data Sets Application Data Sets

21

DB2 Table Refresh StepsProduction DB2

‘Source’

DB2Target

Target DB2

8

9

1

Source Job

7

Target Job

Create target DB2 DDL and table and index spaces if they don’t exist

Perform copy process

Start target table and index spaces

Update identity columns

Object ID translation, data masking and log apply

LISTDEF selection

Stop table space or fuzzy copy

Start, if stopped

2 Verify object compatibility

4

5

6

DB2

3

21

22

Benefits of cloning

All source object are known to the clone Triggers Constraints, Etc.

Stats are identical to the source

Packages are already bound and access paths known

Columns may be masked to protect sensitive data

If it can be done on the source it can be done on the clone

The footprint of the clone can be determined by the user and how the clone is populated

22

23

DB2 Support

DB2 Support DB2 offline DB2 online

Clone from an executing DB2 subsystem Clone from a System Level Backup

DB2 data sharing DB2 data sharing with many to less members DB2 data sharing to non-DB2 data sharing Clone from a System Level Backup

23

24

Using Volume Clone and Rename (VCR) & Fast Tablespace Refresh(FTR)

Smooths the migration path – and helps YOU to avoid surprises and delays during and after

the upgrade process

Provides for creating DB2 subsystem clones, which offer multiple levels of testing environments

Clones are created with minimum skilled resources, minimum time and no interruption to your production systems.

Creating clones of your production and nonproduction environments reduces the risk of impacting critical applications during the migration process.

24

25

Q & A

25


Recommended