Colvin RMAN New Features
- 1. Oracle Database 12c – Recovery Manager New Features Presented by: Andy Colvin December 2, 2013
- 2. About Enkitec — Extensive Oracle Practice – 9 years old — — — — — Education Database Migrations Performance Reviews Remote DBA Support Application Express — Enkitec Extreme Exadata Expo — Irving, TX — June 2-3 — http://www.enkitec.com/e4
- 3. About Me — Working around Oracle since 1999 — Background in systems, network, database — 7 years at Enkitec — Working on Exadata for 4 years — Oracle ACE
- 4. Why Talk About RMAN? — Everybody should use RMAN — It can be quite interesting — I’m a fan of cruel and unusual punishment
- 5. What’s New? — New Security Roles — Pluggable Databases — Run SQL from RMAN — Recover Table — Cross-Platform Restore — Active Duplicate Enhancements
- 6. 12c Includes SYSBACKUP Role — Special role that only has backup privileges — For example, SYSBACKUP does not include SELECT ANY TABLE — Recommended method for connecting to RMAN — rman target ' "user/pass as sysbackup" '
- 7. What’s New? — New Security Roles — Pluggable Databases — Run SQL from RMAN — Recover Table — Cross-Platform Restore — Active Duplicate Enhancements
- 8. Pluggable Databases — Oracle 12c introduces the concept of containers and pluggable databases — Pluggable databases are “virtual” databases — Pluggable databases share memory and redo logs
- 9. Pluggable Database Support — RMAN supports Pluggable Databases — Back up entire Container Database or individual Pluggable Databases — Container database – no changes — PDB point in time recovery
- 10. Containers vs Pluggable Databases — Back up container databases like any “standard” database — Afterwards, open all of your pluggable databases RMAN> 2> { 3> 4> 5> 6> } RMAN> RMAN> RUN SET UNTIL SCN 16747183; RESTORE DATABASE; RECOVER DATABASE; alter database open resetlogs; alter pluggable database all open;
- 11. Pluggable Database Support RMAN> report schema; Report of database schema for database with db_unique_name PLUGGY List of Permanent Datafiles =========================== File Size(MB) Tablespace ---- -------- -------------------1 770 SYSTEM 3 610 SYSAUX 4 60 UNDOTBS1 5 250 PDB$SEED:SYSTEM 6 5 USERS 7 490 PDB$SEED:SYSAUX 8 250 PLUG1:SYSTEM 9 510 PLUG1:SYSAUX 10 5 PLUG1:USERS RB segs ------*** *** *** *** *** *** *** *** *** List of Temporary Files ======================= File Size(MB) Tablespace ---- -------- -------------------1 521 TEMP 2 20 PDB$SEED:TEMP 3 20 PLUG1:TEMP Maxsize(MB) ----------32767 32767 32767 Datafile Name -----------------------+DG/PLUGGY/DATAFILE/FILE +DG/PLUGGY/DATAFILE/FILE +DG/PLUGGY/DATAFILE/FILE +DG/PLUGGY/DATAFILE/FILE +DG/PLUGGY/DATAFILE/FILE +DG/PLUGGY/DATAFILE/FILE +DG/PLUGGY/DATAFILE/FILE +DG/PLUGGY/DATAFILE/FILE +DG/PLUGGY/DATAFILE/FILE Tempfile Name -------------------+DG/PLUGGY/TEMPFILE/FILE +DG/PLUGGY/TEMPFILE/FILE +DG/PLUGGY/TEMPFILE/FILE
- 12. PDB Point In Time Recovery — Recover PDBs individually RMAN> RMAN> 2> { 3> 4> 5> 6> } RMAN> alter pluggable database ERP close; RUN SET UNTIL SCN 1674493; RESTORE PLUGGABLE DATABASE ERP; RECOVER PLUGGABLE DATABASE ERP; alter pluggable database ERP open resetlogs;
- 13. What’s New? — New Security Roles — Pluggable Databases — Run SQL from RMAN — Recover Table — Cross-Platform Restore — Active Duplicate Enhancements
- 14. Running SQL From RMAN — No More “SQL” Tags — Previous versions didn’t support SELECT statements — Useful within backup scripts
- 15. Running SQL From RMAN (2) RMAN> select sysdate from dual; SYSDATE --------11-FEB-13 RMAN> desc dba_pdbs Name Null? --------------------------PDB_ID NOT NULL PDB_NAME NOT NULL DBID NOT NULL CON_UID NOT NULL GUID STATUS CREATION_SCN NOT NULL Type ---------------------------NUMBER VARCHAR2(128) NUMBER NUMBER RAW(16) VARCHAR2(13) NUMBER
- 16. What’s New? — New Security Roles — Pluggable Databases — Run SQL from RMAN — Recover Table — Cross-Platform Restore — Active Duplicate Enhancements
- 17. Recover Table — Recover tables from backups — Useful when you can’t use flashback — Recover tables or table partitions — Remap table names using datapump syntax — Must connect with SYSDBA or SYSBACKUP credentials
- 18. Recover Table (2) RMAN> RECOVER TABLE ACOLVIN1.T 2> UNTIL SCN 1674493 3> AUXILIARY DESTINATION '+DATA' 4> DATAPUMP DESTINATION '+DATA' 5> REMAP TABLE 'ACOLVIN1'.'T':'T_RECOVERED';
- 19. Recover Table - Process — RMAN automatically finds necessary backupsets — Auxiliary database created with backupsets — Data pump export file automatically created — RMAN performs data pump import of the tables to be recovered (optional) — RMAN cleans up after itself, deleting datapump file and temporary instance files
- 20. What’s New? — New Security Roles — Pluggable Databases — Run SQL from RMAN — Recover Table — Cross-Platform Restore — Active Duplicate Enhancements
- 21. Cross-Platform Backup/Restore — Utilizes backupsets (smaller backups to transport) — Great for migrations – “read-only” requirement removed — Allows for shorter downtime when moving across platforms
- 22. Cross-Platform Backup/Restore (2) 1. Perform backup of the source tablespaces 2. Restore to the new target database 3. Take periodic incremental backups and recover on target 4. Place source tablespaces in read-only mode, take final incremental backup and export metadata 5. Restore final incremental backup to target database 6. Import tablespace metadata into target database
- 23. What’s New? — New Security Roles — Pluggable Databases — Run SQL from RMAN — Recover Table — Cross-Platform Restore — Active Duplicate Enhancements
- 24. RMAN Duplicate – Old School — Each channel is assigned a datafile — Files assigned to channels starting with largest first — When a file is finished, the next largest available file is copied — Works perfectly when all datafiles are same size — What if we have different sized datafiles?
- 25. What Does This Mean? — Imagine dozens of datafiles — Ranging from 50GB to 9TB — Allocate too many channels, they will sit idle — Don’t allocate enough channels, wait on largest datafiles ***This is changed in 12c*** — Active duplicate utilizes backupsets
- 26. RMAN Duplicate Channel Allocation Source Database File 1 1TB File 5 10GB File 2 50GB File 6 10GB File 3 10GB File 7 10GB File 4 20GB File 8 10GB RMAN Channel 1 RMAN Channel 2 Target Database
- 27. Active Duplicate Enhancements — Active duplicate defaults to use backupsets — This gives us all of the benefits of backupsets, with active duplicate — PIECE SIZE — COMPRESSION — Empty space moves much faster
- 28. Active Duplicate Enhancements — Example code RMAN> run { 2> allocate channel d1 device type disk; 3> allocate channel d2 device type disk; 4> allocate channel d3 device type disk; 5> allocate auxiliary channel s1 device type disk; 6> duplicate target database for standby 7> from active database piece size 500M; 8> }
- 29. Active Duplicate Enhancements — What the output looks like channel s1: starting datafile backup set restore channel s1: using network backup set from service cloudy channel s1: specifying datafile(s) to restore from backup set channel s1: restoring datafile 00001 to +SMITHERS/windy/datafile/ system.273.807276333 channel s1: restoring section 1 of 2 channel s1: restore complete, elapsed time: 00:00:15 channel s1: starting datafile backup set restore channel s1: using network backup set from service cloudy channel s1: specifying datafile(s) to restore from backup set channel s1: restoring datafile 00001 to +SMITHERS/windy/datafile/ system.273.807276333 channel s1: restoring section 2 of 2
- 30. Active Duplicate Enhancements — Let’s look closer — Using a utility called dstat, we can see what’s going on — dstat -dnyc -D xvda,xvde,xvdf,xvdg -N eth0 -C total — Using a utility called dstat, we can see what’s going on
- 31. Active Duplicate Enhancements --dsk/xvda----dsk/xvde----dsk/xvdf----dsk/xvdg- --net/eth0- ----total-cpu-usage---read writ: read writ: read writ: read writ| recv send||usr sys idl wai hiq siq 0 0 : 0 35M: 0 32M:8192B 34M| 13M 289k|| 1 0 90 9 0 0 0 0 : 0 36M: 0 39M: 0 36M| 109M 2168k|| 1 1 97 1 0 1 0 0 : 0 37M: 32k 37M: 32k 38M| 116M 2299k|| 1 1 97 0 0 empty datafile 1 0 0 : 0 38M: 0 37M: 0 37M| 116M 2345k|| 1 1 97 0 0 1 0 0 : 32k 21M: 96k 22M: 56k 21M| 65M 1311k|| 0 1 97 1 0 0 0 0 : 16k 0 : 16k 16k: 32k 16k| 52B 508B|| 0 0 100 0 0 0 0 0 : 16k 4096B: 32k 0 : 48k 0 | 11k 10k|| 0 0 100 0 0 0 0 4096B: 0 46M: 32k 44M: 16k 44M| 548B 743B|| 1 0 85 14 0 0 0 0 : 0 44M: 0 46M: 24k 46M| 532B 1220B|| 1 0 84 15 0 0 0 16k: 0 26M: 0 29M:8192B 29M|1065k 26k|| 1 0 90 9 0 0 0 0 : 16k 48M: 16k 45M: 32k 42M| 296B 450B|| non-empty14 1 0 85 0 datafile 0 0 0 : 0 46M: 0 46M: 0 46M| 426B 759B|| 1 0 84 15 0 0 0 0 : 0 43M: 0 46M: 0 49M| 584B 434B|| 1 0 85 14 0 0
- 32. A Few More Things… — Recover/Restore files over network — Image copy backups support section size — Storage snapshot optimizations — Specify that duplicated database does not open after duplication
- 33. Questions? Contact Information: Andy Colvin email - andy.colvin@enkitec.com web - http://www.enkitec.com blog- http://blog.oracle-ninja.com twitter - @acolvin