Achieving Continuous Availability for Your Applications with Oracle MAA
0 likes339 views
This document summarizes a presentation on achieving continuous availability for applications with Oracle Maximum Availability Architecture (MAA). The presentation covers: 1. An overview of Transparent Application Continuity (TAC) and how it can be used for planned maintenance and unplanned outages. 2. The high availability requirements needed to ensure business continuity, such as using database services, connection strings, and Active Data Guard. 3. A customer example of how JPMorgan Chase uses Oracle Application Continuity in their Gaia Oracle Service to improve application resilience across the firm.
![• Always use user-defined database services to connect with any database/PDB
• Never use the default service
Oracle CloudWorld Copyright © 2022, Oracle and/or its affiliates
Use Database Services
$ srvctl add service -d <DATABASE NAME> -s <SERVICE NAME> [-pdb <PDB NAME>]
-preferred <INSTANCE NAME, …> -available <INSTANCE NAME, …>
-notification true
-reset_state level1 [21c+]
-failover_restore AUTO
-commit_outcome TRUE
-failovertype AUTO
-replay_init_time 600
-drain_timeout 300
-stopoption IMMEDIATE
-role PRIMARY
10
Application
Continuity (AC)
Workload
Draining
RAC HA](https://image.slidesharecdn.com/achievingcontinuousavailabilityforyourapplicationswithmaa-221202014121-bf32baf1/85/Achieving-Continuous-Availability-for-Your-Applications-with-Oracle-MAA-10-320.jpg)
Achieving Continuous Availability for Your Applications with Oracle MAA
- 1. Achieving Continuous Availability for Your Applications with Oracle MAA October 17–20, 2022 Caesars Forum and The Venetian Las Vegas, NV Oracle CloudWorld Copyright © 2022, Oracle and/or its affiliates
- 2. Oracle CloudWorld Copyright © 2022, Oracle and/or its affiliates Speakers Ian Cookson Senior Principal Product Manager Oracle Clusterware & Application Continuity Oracle Markus Michalewicz Vice President of Product Management Database High Availability, Scalability, and MAA Oracle Doug Burns Vice President Oracle Engineering Lead GTI Platform Services JPMorgan Chase 2
- 3. 1. Why Transparent Application Continuity (TAC) should be your default choice 2. High Availability requirements to ensure Business Continuity 3. How to efficiently use TAC for planned maintenance and unplanned outages 4. What to do when Transparent Application Continuity does not fit 5. Customer Example Oracle CloudWorld Copyright © 2022, Oracle and/or its affiliates Topics Covered in this Presentation 3
- 4. Scale out & Lifecycle Data protection Standardized Reference Architectures for Never-Down Deployments Oracle Maximum Availability Architecture (MAA) Reference architectures Deployment choices HA features, configuration and operational practices Customer insights and expert recommendations Production site Replicated site Replication Generic Systems Engineered Systems BaseDB, ExaDB/ExaCC Autonomous DB Flashback RMAN + ZDLRA Continuous availability Application Continuity Edition-based Redefinition Active replication Active Data Guard RAC Sharding FPP 24/7 GoldenGate Online Redefinition Zero Downtime Migration (ZDM) Bronze Silver Gold Platinum Oracle CloudWorld Copyright © 2022, Oracle and/or its affiliates 4
- 5. Business Continuity • Eliminates downtime for users • In-flight work is preserved • Maintenance is hidden • Predictable performance • Errors visible only if unrecoverable • Builds on top of High Availability Oracle CloudWorld Copyright © 2022, Oracle and/or its affiliates Maximum Availability • Minimizes downtime • In-flight work is lost • Rolling maintenance on DB-level • Predictable runtime performance • Errors may be visible • High Availability building blocks From Maximum Availability to Business Continuity 5
- 6. Transparent Application Continuity (TAC) • For planned maintenance and unplanned outages • Available with Oracle RAC and Active Data Guard • Default on Oracle Autonomous Database Oracle CloudWorld Copyright © 2022, Oracle and/or its affiliates Application Continuity (AC) • For planned maintenance and unplanned outages • Available with Oracle RAC and Active Data Guard • Requires Oracle connection pool One Solution – Two Flavors 6
- 7. • Based on the same HA requirements as AC, TAC covers most SQL use cases: • TAC works with and without Oracle connection pools. • ACchk can be used to determine the protection provided. • Oracle 19c Databases and Clients and later provide best results. • High Availability (HA) requirements include: • Use of Oracle Database services. • Using an optimized TNS connection string. • Draining is part of maintenance operations. • Active Data Guard for Disaster Recovery. • TAC seamlessly improves business continuity on Autonomous Database • Using the same standards, TAC can help improve your application availability, too. Default on Oracle Autonomous Database – Why? Oracle CloudWorld Copyright © 2022, Oracle and/or its affiliates Transparent Application Continuity 7
- 8. And catch applications that use coding practices that prevent safe replays Oracle CloudWorld Copyright © 2022, Oracle and/or its affiliates Determine the Protection Provided by TAC with ACchk Disable Reason ORA-41429: side effect detected 8
- 9. High Availability Requirements Oracle CloudWorld Copyright © 2022, Oracle and/or its affiliates Business Continuity 9
- 10. • Always use user-defined database services to connect with any database/PDB • Never use the default service Oracle CloudWorld Copyright © 2022, Oracle and/or its affiliates Use Database Services $ srvctl add service -d <DATABASE NAME> -s <SERVICE NAME> [-pdb <PDB NAME>] -preferred <INSTANCE NAME, …> -available <INSTANCE NAME, …> -notification true -reset_state level1 [21c+] -failover_restore AUTO -commit_outcome TRUE -failovertype AUTO -replay_init_time 600 -drain_timeout 300 -stopoption IMMEDIATE -role PRIMARY 10 Application Continuity (AC) Workload Draining RAC HA
- 11. T Y P E N O N T L S ( A D B - D) T L S N A M E S TA N D BY ( A D B - D) T L S S TA N D BY ( A D B - D) High priority OLTP 1 tpurgent tpurgent_tls tpurgent_ro tpurgent_ro_tls Typical OLTP 1 tp tp_tls tp_ro tp_ro_tls High priority Reporting 2 high high_tls high_ro high_ro_tls Typical Reporting 2 medium medium_tls medium_ro medium_ro_tls Low priority Reporting 2 low low_tls low_ro low_ro_tls Oracle CloudWorld Copyright © 2022, Oracle and/or its affiliates Example: Autonomous Pre-Created Services 11 1 Transparent Application Continuity enabled by default 2 Use DBMS_APP_CONT_ADMIN.ENABLE_TAC to enable TAC, or Use DBMS_APP_CONT_ADMIN.ENABLE_AC to enable AC
- 12. • Connecting to a custom database service for a “sales application” • Full Description: jdbc:oracle:thin:@ (DESCRIPTION = (CONNECT_TIMEOUT=90) (TRANSPORT_CONNECT_TIMEOUT=3) (RETRY_COUNT=50) (RETRY_DELAY=3) (FAILOVER=ON) (ADDRESS_LIST= (LOAD_BALANCE=on) (ADDRESS=(PROTOCOL=tcp)(HOST=CLOUD-SCANVIP.example.com)(PORT=1522))) (CONNECT_DATA=(SERVICE_NAME=Sales_svc))) Oracle CloudWorld Copyright © 2022, Oracle and/or its affiliates Connection Best Practices 12
- 13. Draining describes • an operation that causes sessions to complete their work on a given instance • to prepare the node or the database for maintenance. Draining is a function of the Oracle Database, supported by Oracle drivers, connection pools, and natively by SQL. Oracle CloudWorld Copyright © 2022, Oracle and/or its affiliates Fast Application Notification (FAN) • Breaks applications out of TCP timeouts. • Notifies clients of status changes on database service-level. • Is automatically configured. Configure Fast Application Notification and use Draining 13
- 14. Oracle CloudWorld Copyright © 2022, Oracle and/or its affiliates Drain… Connect… Failover Time TPM Node 1 Node 2 Drain Timeout (1) Drain Node 1 (2) Connect to Node 2 (4) Replay in-flight requests with Application Continuity (3) Terminate in-flight requests 14
- 15. Planned and Unplanned Outages With Transparent Application Continuity Oracle CloudWorld Copyright © 2022, Oracle and/or its affiliates 15
- 16. • Actions on database level: 1. DB Request interrupted by an error or FAN or drain 2. Session reconnects to the available service 3. Database (DB) Request replays automatically 4. Result from Database Request returned to user • (Transparent) Application Continuity • Fails over within SLA’s • Guarantees commit at most once • Verifies: • Correct database • Correct session state • Correct results Oracle CloudWorld Copyright © 2022, Oracle and/or its affiliates Unplanned Outages and Planned Failover 1 2 3 2 4 Primary Active Data Guard Standby 3 16
- 17. When replay is not enabled Oracle CloudWorld Copyright © 2022, Oracle and/or its affiliates Exclusions Application level • The following is used: • Default database or default PDB service • JDBC OCI (end of life) • Pre-compilers (in development) Remainder of request • Alter system, database (intentionally excluded) • Best effort streaming lobs • XA • Deprecated OCI calls: OCIStmtPrepare, misc api’s • Database links ADG to primary Target database • Different database • DBMS_ROLLING* • Golden Gate • 3rd Party Replication 17
- 18. What to do when Transparent Application Continuity does not fit Oracle CloudWorld Copyright © 2022, Oracle and/or its affiliates 18
- 19. Oracle CloudWorld Copyright © 2022, Oracle and/or its affiliates Oracle Database 19c – Offering Two Flavors TAC – simple apps AC – pooled apps Oracle or 3rd party connection pools with request boundaries Yes, if no unrestorable state* Yes Non-pooled or no request boundaries Yes, if no cursors in fetch and no unrestorable state* No Reenables after COMMIT or other disabling call Yes, if no cursors in fetch and no unrestorable state* No, until request boundary SQL, Transactions & Queries Yes Yes PL/SQL Yes, disables side effects Yes, allows side effects Restores state at start of replay Yes Yes Restores mutables Yes Yes Replays over DBlinks No Yes Replays when client state and results match Yes, server state must also match Yes *Unrestorable state examples: PL/SQL global session state, temp tables, session lobs, OJVM 19
- 20. • Use Transparent Application Continuity (TAC) for applications • Using INSERT, UPDATE, DELETE, SELECT, ALTER SESSION – basic ANSI SQL calls • That may or may not use Oracle connection pools • That can accept that a replay of side effects is disabled automatically • Not requiring transactions over database links to replay • Use Application Continuity (AC) for applications • Using SQL and PL/SQL global session state, temp tables, session LOBs • Can use an Oracle connection pool or other compliant pools (RedHat JBoss) • That need replay of side effects enabled (such as email, audit, logging, file transfer) • Need database link support Oracle CloudWorld Copyright © 2022, Oracle and/or its affiliates Which Flavor to Use 20
- 21. • Packaged applications, from Oracle and other vendors, pose special challenges • What has the vendor tested regarding workload draining and Application Continuity? • Certification and support will constrain what versions and configuration options are available • Often designed for older technology stacks, limiting what can be achieved • Oracle & other packaged applications are adopting draining and Application Continuity • E-Business Suite – support for workload draining with SQL connection tests (under development) • JD Edwards – support for workload draining (TAC replay planned to be available soon) • Fusion Middleware – support for workload draining (GA) and Application Continuity (under development) • PeopleSoft – supports draining and TAC replay • Siebel – has its own replay mechanism • SAP – support for workload draining (TAC replay support planned to be available soon) Oracle CloudWorld Copyright © 2022, Oracle and/or its affiliates Application Continuity for Packaged Applications 21
- 22. Customer Example Oracle CloudWorld Copyright © 2022, Oracle and/or its affiliates 22
- 23. Oracle Application Continuity in Gaia Oracle Service Improving app resilience across the firm by simplifying resilient design for our application engineers Doug Burns Vice President Oracle Engineering Lead GTI Platform Services Madhu Bhimaraju Executive Director Head of Development & Modernization Consumer & Community Banking (Chase)
- 24. 24 Challenges & Solutions Planned Outages • Connection Pool Draining, FAN & TAF • Drain connection pools and allow transactions to complete within application-specific timeframes. • Transactions that don’t complete within the timeout are rolled back. Apps need to detect and replay. Unplanned Outages • In-flight Transactions are always rolled back to preserve database integrity. Apps need to detect and replay. Oracle 19c Transparent Application Continuity • Automatically tracks in-flight transactions and replays them when service is restored. • Zero/Low Dev Effort. Other than connecting to a different Service we provide; possibly defining transaction boundaries; and using a different driver - no detection/recovery code required. • Modern App Design is still relevant. Works best with connection pools and well-designed apps that borrow connections, execute one transaction then return to the pool.
- 25. 25 25 Gaia Oracle Service (GOS) 1 2 3 4 5 Private Oracle Cloud Opinionated. Rapid Provisioning. No-DBA model. Customers Strategic platform for internal Oracle databases 60-70% cloud v. non-cloud in 3 years and on a great trajectory Critical zero-downtime applications e.g. Debit, Payments etc., MAA Closely follows detailed MAA recommendations All Production databases are RAC clusters and have at least one ADG Standby. Security & Hygiene Mandatory application of Release Updates. Opinionated security model. Self-Service … means self-service! Full Repave capabilities, Rolling Patching, Switchover/Failover, Create/Stop/Start Services.
- 26. 26 26 1 2 3 4 5 Key Lessons Application Feature Application, Application Server configuration and great software engineers. For the DBAs in the audience … this requires collaboration. Client Versions Non-negotiable and not a significant issue for us. Testing There is play-testing/PoC and real testing. How do you recreate all possible scenarios? Sense of wonder! Once all the guidance has been followed and requirements met, it’s an amazing feature! Information Sources Google Search is not your friend. Best sources of information are from the Oracle Product Team and are all available at “Oracle Application Continuity”
- 27. Application Continuity Best Practices Thank you Oracle CloudWorld Copyright © 2022, Oracle and/or its affiliates