Upload
walter-ferguson
View
231
Download
1
Tags:
Embed Size (px)
Citation preview
11g(R1/R2) Data guard EnhancementsSuresh Gandhi
2
Agenda
What is DATA GUARD ? Types of Data Guard Databases 11g Release 1 Enhancements 11g Release 2 Enhancements Active Data guard Snapshot Standby Database More Features Rolling Upgrade and Physical Standby Q&A
3
Data Guard Overview
Data Guard
Standby Plus more
High Availability & Disaster Recovery Solution
4
Provides ….
5
Data guard Databases
6
11G R1 Enhancements
Release 1 New
Features
7
R2 - Enhancements
Release 2 New
Features
8
R2- New Features
Release 2 New
Features
9
Active and Snapshot Standby
Traditional Functionality until 10g11g Active DataGuard11g Snapshot Standby
10
Active Data Guard Benefits
11
Turn on Physical Standby to Active StandbyStop redo apply
Database in standby mount
mode
Open readonly
Open mode changed to read
only
Start the redo apply again using
log file (SRL’s)
Status changed to read only apply
12
Turn on Physical Standby to Snapshot StandbyOnce sync check, convert standby to
snapshot
Open database
Read write mode
Need to close database
Switched back to physical standby
13
Additional Info – Active / Snapshot
WHAT OPERATION ALLOWED IN ACTIVE DATA GUARD ? SELECT statements Complex queries Calling of stored procedures Use of database links to redirect DML to Primary Use of stored procedures to call remote stored procedures via database links Use of SET ROLE Use of ALTER SESSION and ALTER SYSTEM No Datatypes restrictions like Logical Standby Additional TEMPFILES can be created for Sorts Flashback Query : SELECT … AS OF <SCN> Additional License
• SOME IMPORTANT POINTS FOR SNAPSHOT STANDBY DATABASE
• PHYSICAL STANDBY DATABASE SHOULD BE IN FLASHBACK MODE
• NO REAL TIME QUERY OR DATA • NO FAST START FAILOVER IN THIS MODE • NO SWITCHOVER/FAILOVER UNTIL
CONVERTED TO PHYSICAL STANDBY • Free of Licence• POINTS
1. Configure the Flash Recovery Area 2. Set the Retention Target
3. Enable Flashback Database
14
10g Equivalent Snapshot commandsPrepare the physical standby database to be activatedSQL> ALTER SYSTEM SET DB_RECOVERY_FILE_DEST_SIZE=20G;SQL> ALTER SYSTEM SET DB_RECOVERY_FILE_DEST='/arch/oradata'
Prepare the primary database to have the physical standby be diverged.
Archive the current log file.
SQL> ALTER SYSTEM ARCHIVE LOG CURRENT;
Cancel Redo Apply and create a guaranteed restore point.SQL> ALTER DATABASE RECOVER MANAGED STANDBY DATABASE CANCEL;SQL> CREATE RESTORE POINT restore_pt GUARANTEE FLASHBACK DATABASE;
Activate the physical standby database.
SQL> ALTER DATABASE ACTIVATE STANDBY DATABASE;SQL> STARTUP MOUNT FORCE;SQL> ALTER DATABASE SET STANDBY DATABASE TO MAXIMIZE PERFORMANCE;SQL> ALTER DATABASE OPEN
Revert the ac4vated database back to a physical standby database.SQL> STARTUP MOUNT FORCE;SQL> FLASHBACK DATABASE TO RESTORE POINT restore_pt;SQL> ALTER DATABASE CONVERT TO PHYSICAL STANDBY;SQL> STARTUP MOUNT FORCE;
15
DATGUARD – Still More Features
16
More Features – RMAN Active database Duplication
RMAN> run {allocate channel prmy1 type disk;allocate channel prmy2 type disk;allocate channel prmy3 type disk;allocate channel prmy4 type disk;allocate auxiliary channel repstby type disk;duplicate target database for standby from active database spfileparameter_value_convert 'emrep','repstby'set db_unique_name='repstby'set db_file_name_convert='C:\','F:\'set log_file_name_convert='C:\','F:\';}
Steps–Create a pfile with only db_name=stby–Create directory structure for datafile locations–Create directory structure for dump locations–Create a password file (must match with target)–Listener & Tnsnames entries for new standby database–Startup nomount with pfile–Run the rman command in other box
Log File Duplicate_Database_log.txt
17
More Features – Redo Compression
Requires Advanced Compression features Compress redo as it is transmitted data Automatically compress the data as it only resolves gaps Helpful and useful in bandwidth constrained locations Syntax:-
LOG_ARCHIVE_DEST_2=‘Service=stby Async compression=ENABLE’
18
More Features – Heterogenous database support
Oracle 10g Dataguard support mixed environments 32bit and 64 bit configuration
Oracle 11G Dataguard supports mixed Windows/Linux in same configuration
Only Physical Standby Same Endianess required on all platforms
Metalink : 413484.1
19
More Features: Lost Write Detection
Due to hardware or storage issues lost writes can happen while transmission and blocks may be stale.
Very hard to diagnose in such of case of corruption
Feature:-– Compares version of Blocks
Between Standby blocks and redo stream Version discrepancy can be either standby or primary
– If Primary Database block corruption is detected Resolved Using standby failover and Restore data consistency
New Initialization parameter– ALTER System set db_lost_write_protect = NONE/TYPICAL/FULL– 5% impact but increase protection
20
More Features: Automatic Block recovery
11g R2 automatically recovers corrupted blocks from standby copies of blocks
Automatic Block Media Recovery also automatically repair corrupted blocks discovered in the physical standby databases
Reduce recovery time in block corruption cases instead of restoring from tape or disks
RMAN command RECOVER BLOCK can be used for manual recovery and its check for all suitable or consistent blocks across several copies of backups or
– Clause can be added RECOVER BLOCK … EXCLUDE STANDBY then wont check in standby
Automatic block recovery is applicable only in physical corruptions like When checksum is invalid Block contains all zeros Block header is fractured
21
More Features: Redo Flush
Yes, Redo can be flushed
Can be done when the primary is in mount mode
To avoid minimal loss of data and allowed failover to be performed without data loss
Syntax:-– ALTER System FLUSH Redo to <target DB Name>
22
More Features: Sync with Primary
New command – ALTER SESSION SYNC WITH PRIMARY
– WAIT FOR STANDBY TO BE SYNCHRONIZED WITH PRIMARY– FAILS IF REDO APPLY OR REDO TRANSPORT IS OFF
REPORT TRANSPORT AND APPLY LAGS
23
More Features: STANDBY_MAX_DATA_DELAY
24
More Features: ROLE Based services
Services can be created and start based on the database ROLE, viz a new addition sys_context (Database ROLE)
In R2 services can be attached via srvctl
Service can failover when using Cluster or Oracle restart
11gR1 to use triggers based on Database role– Example Trigger
role-based-triggers.txt.txt
25
Rolling Upgrade : Transient Logical Standby Physical standby must be converted to Logical Standby and uses Flashback
as fall out plan
Downtime for upgrade is only the SWITCHOVER TIME
26
Finally, Monitoring Views
GAP_STATUS , additional column in v$archive_dest – GAP or NO GAP
V$dataguard_stats for transport/apply lag
V$REDO_DEST_RESP_HISTOGRAM– Useful when need to analyse network latencies by checking transport lag
V$STANDBY_EVENT_HISTOGRAM
Internal 27
Thank you!
Q & A