slide 1

DB2 V10 upgrade experience @ Colruyt

slide 2

DB2 V10 upgrade experience @ Colruyt

• Who are we • Migration plan • Issues • Performance • New features

slide 3

Who are we...

France

Belgium

Other Wholesale & Food service Retail

Luxembourg

Colex

slide 4

Who are we... • DB2 Architecture

DBP1

DBS2

DBT2

DBA1

DBI1

DBP2

DBS1

DBT1

DBA2

DBE1

Legend: • Production

• Development

• Test

• Training

• Acceptence

• DBA Sandbox

Galileï (z196-2817) 6 CPU & 2 zIIP 6251 mips 96 GB

Newton (z196-2817) 5 CPU & 2 zIIP

5300 mips 96 GB

DB2P

DB2S

DB2T

DB2A

DB2I

DB2E

XPC01

XOC01

XAC01

slide 5

Migration plan • New customer since 2008 started in DB2 V9

– No existing migration plans

• Setup DBA sandbox (DB2I) environment (January 2012)

• Pre-migration tasks (January 2012 – February 2012) – Upgrade Third party Tools – Check for necessary V9 PTF’s – Check for incompatibilities / ZPARM changes – DSNTIJPA

• Migration to Conversion Mode (February 2012 – June 2012)

– With preventive maintenance prior to migration DB2P

• Migration to New-Function Mode (August 2012 – October 2012)

slide 6

Migration plan Premigration Tasks

– Migration Third Party Product

• IBM Utility Suite – Log Analysis Tool – Query Monitor Tool – Administration Tool – Tivoli Performance Monitor

• DB2 Connect – Upgraded from V9.1 to V10.1

» V9.7 FP3a required for new functions

• Replication Tools • Windows Client Tools

– Datastudio – AQT / Toad

slide 7

Migration plan Premigration Tasks

– Check for necessary V9 PTF’s

• Current on maintenance (December 2011)) – Check for incompatibilities / ZPARM changes

• A few ZPARM changes were done before migration started – Altered minimal/default values

• Modified DSN3SATH • Deprecated DSNAEXP move to new EXPLAIN privileges (TODO 2013)

– DSNTIJPA • Some PLANs that contained DBRMs • Migration of EXPLAIN tables

– SMS managed Catalog & Directory • Already implemented in V9 through SMS rules • Extended addressability

slide 8

Migration plan SMS Environment Setup

DB2 Catalog

DB2 Directory

Active Log Copy 1

DB2 BSDS 1

Active Log Copy 2

DB2 BSDS 2

User Data

Tool Data

Workfile Database

Archive Logs

GDB2PCP1 GDB2PCP2

CDB2PCP1

DDB2EXT

CDB2PCP2

GDB2PCP3

CDB2PCP3

GDB2PCP4

CDB2PCP4

GDB2PCP5

CDB2PCP5

GDB2PCP7

CDB2PCP7

Backups

Storagegroup

Storageclass

Dataclass

DDB2EXT

slide 9

Migration plan Migration to Conversion Mode

– DSNTIJXZ

• Update default ZPARM input member (DSNTIDxx) • Buffer Pool Settings

– Ensure that no DB2 work is running • Cancelled utilities !!

– DSNTIJMV / DSNTIJRW • New / Modified WLM environments

– DSNTIJEX • Recompile security exit DSN3SATH

– DSNTIJTC • Migrate to Conversion Mode

– Rebind all packages • Rebind with PLANMGMT (EXTENDED)

– Available since DB2 V9 4 hours

slide 10

Migration plan Issues in Conversion Mode

– Issues with IBM Tools

• IBM Administration Tool generating faulty scripts • IBM Query Monitor displaying incorrect elapsed times

– Utility jobs abending with 00C4 • Every time dynamic sort space allocation occurred • Workaround: specify SORTDEVT & SORTNUM • Problem CA-ACF2 interface

– IDLE THREAD TIMEOUT • ABEND=04E REASON=00D3003B • Threads were kept active even after commit

– Are forced out of DB2 (IDHTHOIN = 120 sec) – UK78127 was applied

• Fault in planning – REBIND job ran when member 1 was migrated – WLM choose to run on member 2 – Rebind all fixed the issue

• Migration plan was adapted

slide 11

Migration plan Issues in Conversion Mode

– High CPU consumption in MSTR STC - SRB

• Attributed to internal memory management for statistics • PTF UK80191 resolves the problem

– Currently being implemented – High CPU consumption in MSTR STC – TCB

• Introduced after applying z/OS maintenance • DB2 gathers additional real storage information • z/OS fix UA63988 should reduce the CPU time required to gather this information

– Currently being implemented – Drop of + 70% TCB CPU time in DB2A

slide 12

Migration plan High CPU consumption in MSTR STC – TCB TIME

MSTR - Total Consumption

0

50

100

150

200

250

300

30/0

1/12

6/02

/12

13/0

2/12

20/0

2/12

27/0

2/12

5/03

/12

12/0

3/12

19/0

3/12

26/0

3/12

2/04

/12

9/04

/12

16/0

4/12

23/0

4/12

30/0

4/12

7/05

/12

14/0

5/12

21/0

5/12

28/0

5/12

4/06

/12

11/0

6/12

18/0

6/12

25/0

6/12

2/07

/12

9/07

/12

16/0

7/12

23/0

7/12

30/0

7/12

6/08

/12

13/0

8/12

20/0

8/12

27/0

8/12

3/09

/12

10/0

9/12

17/0

9/12

24/0

9/12

1/10

/12

8/10

/12

15/1

0/12

22/1

0/12

29/1

0/12

5/11

/12

12/1

1/12

19/1

1/12

26/1

1/12

3/12

/12

WEEK

CPU

seco

nds

SSAS_TCB_TIME SSAS_SRB_TIME SSAS_PRE_CP_SRB SSAS_PRE_IIP_SRB TOTAL_CPU

slide 13

Migration plan Migration to New-Function Mode

– DSNTIJEN

• Migrate to Enabling New-Function Mode • Converts the catalog tables

– DSNTIJNF • Migrate to New-Function Mode

– DSNTIJNG • Update DSNHDECP load module

– DSNTIJRT/DSNTIJRV • Install and verify DB2 supplied routines

– DSNTIJCL / DSNTIJMS • ODBC & JDBC support

– Rebind all packages • Rebind with PLANMGMT (EXTENDED)

– REORG Catalog & Directory 2 hours

slide 14

Migration plan Issues in New-Function Mode

– DB2 connect Unlimited Edition V10.1

• SQL1598N An attempt to connect to the database server failed because of a licensing problem. • The db2connectactivate utility activates the license and grants access to SYSIBM.DB2CON_LIC package

– SYSIBM.DB2CON_LIC package: » Will update a local file conlic.bin (if last entry is not older then 7 days)

• CA-ACF2 interface as external security package – Grant to SYSIBM.DB2CON_LIC package was not done by db2connectactivate utility

• After granting EXECUTE on package everything OK

slide 15

Performance • SQL Performance comparison:

– Weekly sum CPU consumption by thread type

– Daily Average CPU consumption per SQL statement by

thread type

– Thread types: • DRDA • CICS • BATCH (PL/I)

slide 16

DRDA performance

0

10.000

20.000

30.000

40.000

50.000

60.000

14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48

WEEK

CPU

seco

nds

0

2

4

6

8

10

12

14

Millio

nsSQ

L STM

T (x 1

.000.0

00)

CLASS2_CPU_AGENT CLASS2_IIP_CPU OCCURENCES

slide 17

DRDA performance (weekdays only)

0,000000

0,000100

0,000200

0,000300

0,000400

0,000500

0,000600

0,000700

2/04/1

2

9/04/1

2

16/04

/12

23/04

/12

30/04

/12

7/05/1

2

14/05

/12

21/05

/12

28/05

/12

4/06/1

2

11/06

/12

18/06

/12

25/06

/12

2/07/1

2

9/07/1

2

16/07

/12

23/07

/12

30/07

/12

6/08/1

2

13/08

/12

20/08

/12

27/08

/12

3/09/1

2

10/09

/12

17/09

/12

24/09

/12

1/10/1

2

8/10/1

2

15/10

/12

22/10

/12

29/10

/12

5/11/1

2

12/11

/12

19/11

/12

26/11

/12

AVG CPU/SQL 30 per. Mov. Avg. (AVG CPU/SQL)

slide 18

DRDA Performance (weekdays only)

• DRDA Performance conclusion: V9

NFM CM NFMAVG_CPU_SQL 0,000174 0,000159 0,000172

% 100,00% 91,13% 98,88%

AVG_IIP_SQL 0,000273 0,000229 0,000253% 100,00% 83,87% 92,81%

V10

93,44%

86,53%

slide 19

CICS performance

0

500

1.000

1.500

2.000

2.500

3.000

3.500

4.000

4.500

14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48

WEEK

CPU

seco

nds

0

100.000

200.000

300.000

400.000

500.000

600.000

700.000

800.000

900.000

1.000.000

SQL S

TMT (

x 1.00

0.000

)

CLASS2_CPU_AGENT OCCURENCES

slide 20

CICS performance (weekdays only)

0,000000

0,000200

0,000400

0,000600

0,000800

0,001000

0,001200

2/04/1

2

9/04/1

2

16/04

/12

23/04

/12

30/04

/12

7/05/1

2

14/05

/12

21/05

/12

28/05

/12

4/06/1

2

11/06

/12

18/06

/12

25/06

/12

2/07/1

2

9/07/1

2

16/07

/12

23/07

/12

30/07

/12

6/08/1

2

13/08

/12

20/08

/12

27/08

/12

3/09/1

2

10/09

/12

17/09

/12

24/09

/12

1/10/1

2

8/10/1

2

15/10

/12

22/10

/12

29/10

/12

5/11/1

2

12/11

/12

19/11

/12

26/11

/12

3/12/1

2

AVG CPU/STMT 30 per. Mov. Avg. (AVG CPU/STMT)

slide 21

CICS Performance (weekdays only)

• CICS Performance conclusion: V9

NFM CM NFMAVG_CPU_SQL 0,000766 0,000721 0,000774

% 100,00% 94,13% 101,04%

V10

96,12%

slide 22

BATCH performance

0

10.000

20.000

30.000

40.000

50.000

60.000

70.000

80.000

10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48

WEEK

CPU

seco

nds

0

500

1.000

1.500

2.000

2.500

3.000

Millio

nsSQ

L STM

T (x 1

.000.0

00)

CLASS7_CPU_AGENT CLASS7_IIP_CPU SQL_STMTS_ISSUED

slide 23

BATCH performance

0,000000

0,000005

0,000010

0,000015

0,000020

0,000025

0,000030

0,000035

0,000040

0,000045

0,000050

2/03/12

12/03/12

22/03/121/04/12

11/04/12

21/04/121/05/12

11/05/12

21/05/12

31/05/12

10/06/12

20/06/12

30/06/12

10/07/12

20/07/12

30/07/129/08/12

19/08/12

29/08/128/09/12

18/09/12

28/09/128/10/12

18/10/12

28/10/127/11/12

17/11/12

27/11/120

500.000.000

1.000.000.000

1.500.000.000

2.000.000.000

2.500.000.000

3.000.000.000

3.500.000.000

4.000.000.000

4.500.000.000

5.000.000.000

AVG CPU/STMT SQL_STMTS_ISSUED 30 per. Mov. Avg. (AVG CPU/STMT)

slide 24

BATCH Performance

• BATCH Performance conclusion:

V9NFM CM NFM

AVG_CPU_SQL 0,000022 0,000180 0,000012% 100,00% 83,54% 57,36%

V10

75,80%

slide 25

Subsystem Performance - MSTR

• MSTR CPU consumption increased: – After CM:

• Due to DB2 internal memory managemnt process – PREEMPT SRB – Fixed in UK80191 currently being applied

• +45% for TOTAL_MSTR_CPU

– After Z/OS maintenance: (25/08/2012) • z/OS PTF UA60825

– Fixed in UA63988 applied in XOC01

• +300% in TCB_MSTR_CPU

slide 26

Subsystem Performance - MSTR

MSTR - Total Consumption DB2P

0

500

1000

1500

2000

2500

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48

WEEK

CP

U s

econ

ds

SSAS_TCB_TIME SSAS_SRB_TIME SSAS_PRE_CP_SRB SSAS_PRE_IIP_SRB

slide 27

Subsystem Performance – DBM1

• DIST CPU consumption moved to zIIP: – DBM1_SRB_CPU

• -50%

– DBM1_SRB_IIP • +200%

– DBM1_TOTAL_CPU

• Netto saving of 40%

slide 28

Subsystem Performance – DBM1

DBM1 - Total Consumption DB2P

0

5000

10000

15000

20000

25000

30000

35000

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48

WEEK

CPU

seco

nds

DSAS_TCB_TIME DSAS_SRB_TIME DSAS_PRE_CP_SRB DSAS_PRE_IIP_SRB

slide 29

Subsystem Performance – DIST

DDF - Total Consumption DB2P

0

10000

20000

30000

40000

50000

60000

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48

WEEK

CPU

seco

nds

DDF_TCB_TIME DDF_SRB_TIME DDF_PRE_CP_SRB DDF_PRE_IIP_SRB

slide 30

RUNSTATS CPU/IIP TIME DB2 RUNSTATS

0

200

400

600

800

1.000

1.200

1.400

1.600

14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48

WEEK

CP

U S

EC

0

2000

4000

6000

8000

10000

12000

OC

CU

RR

EN

CE

S

SUM_CL1_CPU SUM_CL1_IIP OCCURRENCES

slide 31

New-features

• REBIND options APREUSE/APCOMPARE • New security privileges • Temporal Data • 1MB FRAME size • Index INCLUDE • …

slide 32

Questions ?