Transcript Slide 1

Oracle Database 11g: Get There With Less Effort, Less Risk, Less
Infrastructure, and More Reliability
Alan Levine
Oracle Enterprise Architect, Oracle Consulting
The following is intended to outline our general
product direction. It is intended for information
purposes only, and may not be incorporated into any
contract. It is not a commitment to deliver any
material, code, or functionality, and should not be
relied upon in making purchasing decisions.
The development, release, and timing of any
features or functionality described for Oracle’s
products remains at the sole discretion of Oracle.
Session Agenda
•
•
•
•
•
•
Introduction
Dissecting the Upgrade Lifecycle
Pre Upgrade Phase
Upgrade Phase
Post Upgrade Phase
Summary
<Insert Picture Here>
How Do Customers Get to Oracle Database 11g?
• New Oracle Database deployment
• Existing Oracle Database version upgrade such as
10.2.0.4 to 11.1.0.7 (same platform)
• Platform Migration
– UNIX to Linux
– Hardware refresh (same operating system)
• Combined version upgrade and platform migration
• Migration to Oracle Database 11g from a non-Oracle
database
Common Goals for Database Upgrades, Migrations,
and New Deployments
Reduced
Effort
Lower
Risk
Less
Infrastructure
Greater
Reliability
Session Agenda
•
•
•
•
•
•
•
Introduction
Dissecting the Upgrade Lifecycle
Pre Upgrade Phase
Upgrade Phase
Post Upgrade Phase
Infrastructure Optimization
Summary
<Insert Picture Here>
Database Upgrade Lifecycle
Pre Upgrade
Upgrade
Planning
Upgrade
Testing
11g
Testing
Upgrade
Dry Run
Production
Migration
Post
Upgrade
Break
- Fix
Normal
Operations
Database Upgrade Lifecycle
Pre Upgrade
Upgrade
Planning
Upgrade
Testing
•Business Requirements
•Technical Requirements
•Prepare Alternatives
•Prepare Test Plans
11g
Testing
Upgrade
Dry Run
Production
Migration
Post
Upgrade
Break
- Fix
Normal
Operations
Database Upgrade Lifecycle
Pre Upgrade
Upgrade
Planning
Upgrade
Testing
11g
Testing
•Develop Upgrade Approach
•Validate Upgrade Approach
•Perfect Upgrade Approach
•Rollback Testing
Upgrade
Dry Run
Production
Migration
Post
Upgrade
Break
- Fix
Normal
Operations
Database Upgrade Lifecycle
Pre Upgrade
Upgrade
Planning
Upgrade
Testing
11g
Testing
•Functional Testing
•Performance Testing
•Integration Testing
•Acceptance Testing
Upgrade
Dry Run
Production
Migration
Post
Upgrade
Break
- Fix
Normal
Operations
Database Upgrade Lifecycle
Pre Upgrade
Upgrade
Planning
Upgrade
Testing
Upgrade
11g
Testing
Dry Run
•Refresh Environments
•Re-Run Upgrade
Production
Migration
Post
Upgrade
Break
- Fix
Normal
Operations
Database Upgrade Lifecycle
Pre Upgrade
Upgrade
Planning
Upgrade
Testing
11g
Testing
Upgrade
Dry Run
Production
Migration
•Upgrade
•(Rollback)
Post
Upgrade
Break
- Fix
Normal
Operations
Database Upgrade Lifecycle
Pre Upgrade
Upgrade
Planning
Upgrade
Testing
11g
Testing
Upgrade
Dry Run
Production
Migration
Post
Upgrade
Break
- Fix
•Monitor
•Troubleshoot
•(Rollback)
Normal
Operations
Database Upgrade Lifecycle
Pre Upgrade
Upgrade
Planning
Upgrade
Testing
11g
Testing
Upgrade
Dry Run
Production
Migration
Post
Upgrade
Break
- Fix
Normal
Operations
•Monitor
•Troubleshoot
Session Agenda
•
•
•
•
•
•
•
Introduction
Dissecting the Upgrade Lifecycle
Pre Upgrade Phase
Upgrade Phase
Post Upgrade Phase
Infrastructure Optimization
Summary
<Insert Picture Here>
Pre Upgrade Phase
Upgrade Planning Strategies
• Business Requirements
– Understand SLA’s
– Determine available migration windows
– Identify high risk business functions
• Technical Requirements
– Availability of swing servers and storage
– Certification of components for target platform
– Availability of test environments, tools, and
scripts
Pre Upgrade Phase
Upgrade Planning Strategies (cont.)
• Prepare Alternatives
Lower Risk
– Develop strategies that meet business needs
balanced by availability of supporting
infrastructure
•Understand requirements
– Include rollback planning
•Avoid complexity
– Validate key technical assumptions
– Select go-forward approach
• Prepare 11g Test Plans
– Develop in advance to avoid oversights
Reduced Effort
Greater Reliability
•Testing as part of the process, not
an afterthought
•Robust rollback strategy
Selecting a Rollback Strategy
• Restore from backups
• Clone production, upgrade clone, use original production
database for rollback
• Clone production, upgrade production, use clone for rollback
• Export/import to populate a new Oracle 11g database
– Original source database can be used for rollback
– Replication can be used to extend the rollback window beyond the
production 11g cutover
• Data Guard
– Upgrade primary, leave standby untouched until go-live; or,
– Deploy 2nd standby for rollback purposes, upgrade both primary and
standby to avoid compromise of RTO / RPO during upgrade
Pre Upgrade Phase
Upgrade Testing Strategies
• Develop Upgrade Approach (Upgrade Trial 1)
– Perform full upgrade with documentation of all
steps including timing
• Validate Upgrade Approach (Upgrade Trial 2)
– Execute upgrade per the documented plan and
update timing of all steps
– Any deviations from the documented plan
require an additional iteration
– Significant time deviations between Trial 1 and
Trial 2 require further investigation
Pre Upgrade Phase
Upgrade Testing Strategies (cont.)
• Perfect Upgrade Approach (Upgrade Trials
3..n)
– Execute upgrade per the documented plan
– Validate timing of all steps
– No deviations from the plan
• Rollback Testing
– Test the rollback strategy at multiple points in
time during the upgrade process
Lower Risk
•Test your rollback strategy
Greater Reliability
•Perform multiple iterations until the
process is repeatable
Pre Upgrade Phase
Oracle Database 11g Testing Strategies
• Functional Testing
– Leverage automated tools for a repeatable
process
– Insure complete coverage of all critical
business functions and adequate coverage of
lower priority modules
• Performance Testing
– Test production sized workloads
– Try and find the breaking points
– Test real-world combinations of competing
workloads
– Deploy Oracle Real Application Testing (RAT)
• Test with true production workloads using
Database Capture and Database Replay
• Leverage SQL Performance Analyzer (SPA)
performance regression testing
Oracle Real Application Testing
Database Replay
SQL Performance Analyzer
• Replay actual production database
workload in test environment
• Identify, analyze and fix potential
instabilities before making changes to
production
• Enables identification of SQL performance
regressions before end-users can be
impacted
• SPA can help with any change that impacts
SQL execution plan
• Capture Workload in Production
– Capture full production workload with real
load, timing & concurrency characteristics
– Move the captured workload to test system
• Replay Workload in Test
– Make the desired changes in test system
– Replay workload with full production
characteristics
– Honor commit ordering
• Analyze & Report
– Errors
– Data divergence
– Performance divergence
–
–
–
DB upgrades
Optimizer statistics refresh
New indexes, Materialized Views, Partitions,
etc.
• Automates SQL performance tracking of
hundreds of thousands of SQL statements
– impossible to do manually
• Captures SQL workload with low overhead
• Integrated with SQL Tuning Advisor and
SQL Plan Baselines for regression
remediation
Pre Upgrade Phase
Oracle Database 11g Testing Strategies (cont.)
Lower Risk
• Integration Testing
–
–
–
–
Validate all upstream and downstream feeds
Batch jobs and utility scripts
Backup and recovery procedures
Data Guard Switchover and Failover
•Let testing results drive the go-live
date, not vice-versa
Reduced Effort
•Leverage DB Replay and SPA to
automate testing efforts
• User Acceptance Testing
– Deploy the Grid Control Service Level
Management Pack to run synthetic
transactions and provide an objective view of
user workloads
– Deploy Oracle Real User Experience Insight
(RUEI) for non-intrusive views into actual user
performance
Less Infrastructure
•RAT can reduce database testing
infrastructure needs without
compromising testing coverage
Greater Reliability
•Robust “real world” testing scenarios
for a more reliable post-upgrade
experience
Pre Upgrade Mistakes to Avoid
• “We will achieve our go-live date no matter what!”
• “We upgraded four similar databases so we only needed to
performance test one of them.”
• “We ran out of time for performance testing but we should
have enough hardware capacity to mitigate the risk in
production.”
• “We don’t have time for code remediation. We’ll add BIND
VARIABLES later, after we are on 11g.”
• “We did all our development and testing on 10g but it
should be OK to go-live on 11g.”
Session Agenda
•
•
•
•
•
•
•
Introduction
Dissecting the Upgrade Lifecycle
Pre Upgrade Phase
Upgrade Phase
Post Upgrade Phase
Infrastructure Optimization
Summary
<Insert Picture Here>
Upgrade Phase
Dry Run Strategies
• Refresh Test Environments from Production
– Provides a clean environment for validation
of timing
– Helps identify undocumented application
changes
• Re-Run Upgrade
– Should be able to replay the upgrade script
exactly as documented
– Timing of upgrade steps should match
documented times
– Don’t skip this step, it’s the last opportunity to
catch an issue
Lower Risk
•Perform a final dry-run immediately
prior to go-live
Greater Reliability
•You are more likely to have a
successful production upgrade if you
can complete the dry-run without
incident
Upgrade Phase
Production Upgrade Strategies
• Upgrade
Lower Risk
– Don’t deviate from your documented
procedures
•Stick to the plan
– Perform an appropriate application
“smoke test”
Reduced Effort
•It takes less effort to fix a problem
prior to go-live vs. post go-live
– Allow time for user validation
• (Rollback)
Greater Reliability
– Don’t be afraid to rollback if needed
•Wishful thinking will not resolve a
problem once you are in production,
get the kinks out before then
When in doubt, you should back it out!
Upgrade Phase Mistakes to Avoid
• “We don’t need a dry-run. If we have an issue during
the production upgrade we will rollback.”
• “We had a data integrity issue during the migration
but decided we would just fix it after go-live.”
Session Agenda
•
•
•
•
•
•
•
Introduction
Dissecting the Upgrade Lifecycle
Pre Upgrade Phase
Upgrade Phase
Post Upgrade Phase
Infrastructure Optimization
Summary
<Insert Picture Here>
Post Upgrade Phase
Break Fix
• Monitor
– More focused monitoring needs to be in
place for a full [week/month/quarter]
– Measure both the user experience as well
as resource utilization with tools such as
Oracle Grid Control and Oracle RUEI
– Establish new monitoring baselines for key
performance metrics and resource
utilization
• Troubleshoot
– Use a major upgrade as an opportunity to
modernize use of tools
• (Rollback)
– Post go-live rollback strategies are usually
complex and expensive to achieve
Lower Risk
•Better testing up front reduces the
need to consider a post upgrade
rollback strategy
Reduced Effort
•Leverage tools proactively and take
advantage of automated monitoring
and notification
Greater Reliability
•Keep critical resources available for
rapid resolution of issues
Post Go-Live Rollback Options
• Maintain two environments in parallel for as long as mandated
– Feed data to two environments simultaneously
– Increases both hardware and software footprint
– Data divergence due to application changes reduces effectiveness over
time
• Replicate changes from 11g production back to the pre-11g
database environment
– Requires additional pre-upgrade testing to validate the data replication
process
– Increases both hardware and software footprint
• Downgrade the database to the pre 11g version
– Requires additional testing cycles
– Executing a rollback can be time consuming and exceed SLA’s for
database availability
Post Upgrade Phase
Normal Operations
• Monitor
– Continue to leverage tools such as Oracle
Grid Control, the Service Level
Management Pack, and Oracle RUEI
– Fine tune notification thresholds for
deviations from key performance baselines
• Troubleshoot
– Add Oracle Grid Control tools to standard
operating procedures for troubleshooting
• (Rollback)
– No longer an option
Less Risk
•Objective monitoring tools take the
guesswork out of troubleshooting
efforts
Less Effort
•Automation can help proactively
identify issues before users declare a
problem exists
Greater Reliability
•Maintain your testing environment
for future upgrade and patch testing
Mitigating Subjective Performance Feedback
Use Objective Means to Measure the User Experience
• Proactively monitoring the user experience helps to eliminate subjective
performance measurements and avoids unnecessary troubleshooting efforts
• Oracle Grid Control Service Level Management Pack
– Define, deploy, monitor, and report on transaction performance from multiple locations
within your organization
• Oracle Real User Experience Insight (RUEI)
– Monitor and analyze real user activity without needing to create synthetic transactions
Remote
Local
Typical day with impact
Monitoring shows a ‘network event’
of remote site bandwidth limitation
Impacting the remote site
Post Upgrade Phase Mistakes to Avoid
• “We are too busy to sit around all day and watch the
Oracle Grid Control GUI console.”
• “The upgrade went so smoothly that we decided to
double our production workload the next week.”
Session Agenda
•
•
•
•
•
•
•
Introduction
Dissecting the Upgrade Lifecycle
Pre Upgrade Phase
Upgrade Phase
Post Upgrade Phase
Infrastructure Optimization
Summary
<Insert Picture Here>
Server Optimization Strategies
• Goals
– Increase server utilization
– Reduce server footprint
• Strategies
– Use virtualization with OracleVM to increase server utilization
– Share DR with non-production instances
– Leverage Grid Architecture to consolidate databases on fewer
servers
Storage Optimization Strategies
• Goals
– Reduce amount of un-utilized storage
– Reduce overall storage footprint
– Reduce storage costs
• Strategies
– Consolidate database storage with Oracle 11g Automatic
Storage Management (ASM)
– Deploy Oracle Advanced Compression Option to reduce
database storage footprint
– Leverage an Information Lifecycle Management (ILM)
strategy to reduce costs
Disaster Recovery Optimization Strategies
• Goals
– Obtain better return on investment from DR hardware investments
• Strategies
– Co-locate Data Guard standby databases with development and
test instances
• Consider OracleVM if isolation of production environments is
important
– Leverage Snapshot Standby database
• Avoids frequent production to test clones
• Reduces storage footprint
– Use Active Data Guard
• Put DR server capacity to use offloading Read Only query
activity from production
Infrastructure Optimization and
Enterprise Architecture
• Incorporate Infrastructure Optimization as
part of your Enterprise Architecture (EA)
program to better align business needs
with IT challenges
• Oracle Enterprise Architecture
Framework (OEAF) and Oracle
Architecture Development Process
(OADP) support a practical and
iterative approach to EA that delivers
“just enough architecture” based on
customer’s needs
Summary
• Develop upgrade strategies that meet your business
requirements
• Perform planning and testing proportional to the business impact
of an outage
• Leverage tools appropriately both pre and post upgrade to
reduce effort, risk, and cost
• Prepare and test your rollback strategy
• Incorporate regular patching and upgrades into your standard
operating procedures to reduce risks of running on unsupported
releases
• Use the potential benefits of Oracle 11g R2 and Infrastructure
Optimization strategies to make the business case to upgrade
For More Info On Oracle Consulting
• At Oracle OpenWorld:
– Oracle Consulting Center – Moscone South Room 101
• Sign up for 1:1 consultations about Oracle solutions
– Attend Breakout sessions – Search by tag ‘Consulting’
– On DemoGrounds – Moscone West 134/ 135
• On Internet: Oracle.com/consulting
• Email: [email protected]
Featured 2009 OpenWorld Sessions
Plan
Transform
S311839
Optimize
Measure
S311909
www.oracle.com/consulting
© 2008 Oracle Corporation – Proprietary and Confidential