Deep Dive: More Oracle Data Pump Performance Tips and Tricks
1 like1,509 views
The document provides an in-depth look at performance tips and tricks for Oracle Data Pump, emphasizing best practices and important fixes for optimal functioning. It discusses hardware and database configurations, the impact of various settings on performance, and the benefits of using specific patches. Key recommendations include using parallel processing, managing database statistics efficiently, and avoiding unnecessary connections to enhance Data Pump operations.
Deep Dive: More Oracle Data Pump Performance Tips and Tricks
- 1. Copyright © 2015, Oracle and/or its affiliates. All rights reserved. | eep Dive: More Oracle Data Pump erformance Tips and Tricks n Gagne sulIng Member of Technical Staff base UIliIes e CorporaIon Deep Dive: Data Pump Performance Tips & Tricks Roy F. Swonger Sr. Director & Product Manager Database Upgrade & UIliIes Oracle CorporaIon
- 2. Copyright © 2015, Oracle and/or its affiliates. All rights reserved. | e Harbor Statement following is intended to outline our general product direcIon. It is intended for rmaIon purposes only, and may not be incorporated into any contract. It is not a mmitment to deliver any material, code, or funcIonality, and should not be relied u making purchasing decisions. The development, release, and Iming of any features cIonality described for Oracle’s products remains at the sole discreIon of Oracle. Deep Dive: Data Pump Performance Tips & Tricks
- 3. Copyright © 2015, Oracle and/or its affiliates. All rights reserved. | ore Oracle Data Pump Performance Tips & Tricks Deep Dive: Data Pump Performance Tips & Tricks IntroducIon Recap: Performance Best PracIce Important Performance Fixes Futures
- 4. Copyright © 2015, Oracle and/or its affiliates. All rights reserved. | ore Oracle Data Pump Performance Tips & Tricks Deep Dive: Data Pump Performance Tips & Tricks IntroducIon Recap: Performance Best PracIce Important Performance Fixes Futures
- 5. Copyright © 2015, Oracle and/or its affiliates. All rights reserved. | > whoami Deep Dive: Data Pump Performance Tips & Tricks 16 years w/Oracle Previously with DEC OpenVMS – 10 years -‐ Exp/imp Jan 2000 -‐ Metadata May 2001 -‐ Data Pump Mar 2002 -‐ Currently responsible for -‐ Worker process – Mar 2002 – Day 1 -‐ Transportable Tablespace – Apr 2005 -‐ MCP process – May 2007 -‐ Export/Import staIsIcs – Mar 2000 n Gagne
- 6. Copyright © 2015, Oracle and/or its affiliates. All rights reserved. | > whoami years w/Oracle eviously with DEC Rdb + years managing ta Pump tabase Upgrade QL*Loader ansportable Tablespaces Upgrade and Migrate to Oracle 12c -‐ Live and Uncensored Swonger
- 7. Copyright © 2015, Oracle and/or its affiliates. All rights reserved. | roducIon: Test Hardware ystem: Sun Fire X4170 M3 PU Intel x86-‐64 16 total cores x 2.9 Ghz Two-‐way hyperthreaded: CPU_COUNT=32 orage: 4.7 TB in RAID 0 (striped across two volumes) AM: 256 GB S: Oracle Linux 6.4 Deep Dive: Data Pump Performance Tips & Tricks Good amount of CPU for parallel processing Decent Storage Space bu limited performance wit all files on one device Seems like a lot un>l you try running 4 databases with lots of parallelism…
- 8. Copyright © 2015, Oracle and/or its affiliates. All rights reserved. | roducIon: Test Database racle E-‐Business Suite database Lots of metadata • 297 users • 4213 system grants • 12455 sequences • 51064 synonyms • 4364 types • 33164 tables • 375,000+ object grants • 53,196 indexes • 53,217 packages • 7082 triggers – Not a huge amount of data… • Started with ~50GB of data • Built larger tables for some cases, up to ~500GB – …but lots of metadata • Heavy stress on dicIonary performance – Export was from 11.1.0.7 into 16 dumpfiles • Natural limit to parallelism based on # of fi Deep Dive: Data Pump Performance Tips & Tricks
- 9. Copyright © 2015, Oracle and/or its affiliates. All rights reserved. | seline Performance mport into 11.2.0.3 Connected as SYS Simple Parfile: irectory=EBSIMP umpfile=aexp%U.dmp ogfile=full_4_11203.log ULL=Y XCLUDE=TABLESPACE arallel=4 etrics=y an you spot any problems? Deep Dive: Data Pump Performance Tips & Tricks Import time: 7h 53m 52s
- 10. Copyright © 2015, Oracle and/or its affiliates. All rights reserved. | ore Oracle Data Pump Performance Tips & Tricks Deep Dive: Data Pump Performance Tips & Tricks IntroducIon Recap: Performance Best PracIce Important Performance Fixes Futures
- 11. Copyright © 2015, Oracle and/or its affiliates. All rights reserved. | ta Pump Best PracIces ormance Tips se PARALLEL=n Typically n = 2x <number of CPU cores> XCLUDE=STATISTICS on export nable parallel import of indexes: apply patch for bug 21539301 Available for 11.2.0.4 and 12.1.0.2 ew feature in 12c: TRANSFORM=DISABLE_ARCHIVE_LOGGING:Y Apply patch for bug 20778442 Deep Dive: Data Pump Performance Tips & Tricks
- 12. Copyright © 2015, Oracle and/or its affiliates. All rights reserved. | eed up the EsImate Phase What is the EsImate hase? STIMATE=BLOCKS vs STIMATE=STATISTICS onclusion: the more data ou have, the bigger the enefit received from STIMATE=STATISTICS (But make sure you have accurate staIsIcs!) Deep Dive: Data Pump Performance Tips & Tricks 0 20 40 60 80 100 120 140 160 180 200 0 50 100 150 200 250 300 350 400 Es>mate Phase (seconds) Data Size (GB) Comparison of Es>mate Phase Methods Blo Sta
- 13. Copyright © 2015, Oracle and/or its affiliates. All rights reserved. | clude StaIsIcs from Export it faster to EXCLUDE=STATISTICS and build them awer import, or Export and import staIsIcs? est Results: expdp of staIsIcs: 19 minutes impdp of staIsIcs: 83 minutes exec dbms_stats.gather_database_stats(): 82 minutes onclusion: EXCLUDE=STATISTICS saves ~25% of total Ime (plus 150 MB of dumpfile spa Deep Dive: Data Pump Performance Tips & Tricks e Ime and disk space
- 14. Copyright © 2015, Oracle and/or its affiliates. All rights reserved. | NOT Connect as SYS or SYSDBA We do not export SYS-‐owned objects anyway, so SYS is really not ecessary for data pump operaIons onnecIng as SYS (or SYSDBA) also gives worse performance! Import of table using PARALLEL=8 from 8-‐dumpfile export… Connected as schema user GL: mported "GL"."GL_BALANCES” 13.82 GB 147526320 rows in 83 seconds Same user connected as SYSDBA: mported "GL"."GL_BALANCES” 13.82 GB 147526320 rows in 206 seconds onclusion: Use roles and privileges (e.g. DATAPUMP_IMP_FULL_DATABASE) not super use Deep Dive: Data Pump Performance Tips & Tricks roles and privileges to do what you need
- 15. Copyright © 2015, Oracle and/or its affiliates. All rights reserved. | ore Oracle Data Pump Performance Tips & Tricks Deep Dive: Data Pump Performance Tips & Tricks IntroducIon Recap: Performance Best PracIce Important Performance Fixes Futures
- 16. Copyright © 2015, Oracle and/or its affiliates. All rights reserved. | portant Performance Fixes atch for bug number 21539301 Old behavior: build one index at a Ime with PARALLEL n keyword New behavior: build n indexes at a Ime, each without parallelism est results for 53190 indexes 12.1.0.2.0 without patch: • PARALLEL=4: 5681 seconds (total import 4:14:23) • PARALLEL=8: 5504 seconds (total import 3:30:14) 12.1.0.2.3 with patch: • PARALLEL=4: 1308 seconds (total import 2:38:29) • PARALLEL=8: 719 seconds (total import 2:03:11) • PARALLEL=32: 347 seconds (total import 1:44:41) Deep Dive: Data Pump Performance Tips & Tricks allel Import for Indexes in 11.2.0.4/12.1.0.2
- 17. Copyright © 2015, Oracle and/or its affiliates. All rights reserved. | portant Performance Fixes ackground: package bodies have always been loaded in parallel But, loading lots of large packages (e.g. XLA packages in EBS) was sIll really slow ug 16018315 filed, fixed in 2012 (included in 11.2.0.4, 12.1.0.2) ow much improvement can a single bug fix make, really? For 52093 Package Bodies 11.2.0.3: 22179 seconds 11.2.0.4/12.1.0.2: 2810 seconds emark: you aren’t sIll running 11.2.0.3 or 12.1.0.1…are you? Deep Dive: Data Pump Performance Tips & Tricks kage Body Import Improvements
- 18. Copyright © 2015, Oracle and/or its affiliates. All rights reserved. | t’s Put It All Together Deep Dive: Data Pump Performance Tips & Tricks 0:00:00 1:12:00 2:24:00 3:36:00 4:48:00 6:00:00 7:12:00 8:24:00 Import Time Improves, Step by Step 11.2.0.3 parallel=4 baseline 11.2.0.4 parallel=4 with package body fix 11.2.0.4 parallel=4 excluding sta>s>cs 12.1.0.2 parallel=8 12.1.0.2 parallel=8 w/index patch 12.1.0.2 parallel=32 w/index patch
- 19. Copyright © 2015, Oracle and/or its affiliates. All rights reserved. | ore Oracle Data Pump Performance Tips & Tricks Deep Dive: Data Pump Performance Tips & Tricks IntroducIon Recap: Performance Best PracIce Important Performance Fixes Futures
- 20. Copyright © 2015, Oracle and/or its affiliates. All rights reserved. | Deep Dive: Data Pump Performance Tips & Tricks
- 21. Copyright © 2015, Oracle and/or its affiliates. All rights reserved. | e Harbor Statement preceding is intended to outline our general product direcIon. It is intended for rmaIon purposes only, and may not be incorporated into any contract. It is not a mmitment to deliver any material, code, or funcIonality, and should not be relied u making purchasing decisions. The development, release, and Iming of any features cIonality described for Oracle’s products remains at the sole discreIon of Oracle. Deep Dive: Data Pump Performance Tips & Tricks