Upload
others
View
18
Download
0
Embed Size (px)
Citation preview
Copyright © 2015, Oracle and/or its affiliates. All rights reserved. |
Maximum Availability Architecture (MAA) Best Practices for Oracle Database 12c
Joseph Meeks Senior Director, Product Management, Oracle
Michael Smith Consulting Member of Technical Staff, Oracle
Sandip Davda Senior Database Administrator, LinkedIn
Copyright © 2015, Oracle and/or its affiliates. All rights reserved. |
Note: Many of the following slides contain links to MAA Best Practices or My Oracle Support Notes that provide additional details. Click on the text box in the top right-hand corner of the slides to be taken directly to the paper or note that is referenced
Copyright © 2015, Oracle and/or its affiliates. All rights reserved. |
Program Agenda
Oracle MAA
MAA Best Practices
Wrap / Resources
1
2
3
3
4
Copyright © 2015, Oracle and/or its affiliates. All rights reserved. |
Percentage of all unplanned outages due to people and process issues 80%
Average cost of downtime per hour
$10M
Average resolution time per outage
4
$350K
Source: Ponemon Institute, KPMG Tech Risk Radar, Forrester Research
Impact of Database Downtime
Average cost of unplanned data center outage or disaster
3 hours
Copyright © 2015, Oracle and/or its affiliates. All rights reserved. |
Aligns Oracle Solutions with Customer Business Objectives
Prod/Departmental
Business Critical
Development & Test
Mission Critical
• Oracle Database • Backups
Bronze + • Database Clusters
5
Oracle MAA Reference Architecture
Silver + • Simple Physical
Replication
Gold + • Advanced Logical Replication
GOLD
BRONZE SILVER
PLATINUM
www.oracle.com/goto/maa
MAA Best Practices: Reference Architectures
for DBaaS ( embedded url)
Copyright © 2015, Oracle and/or its affiliates. All rights reserved. |
Oracle MAA for Hybrid Cloud
6
Oracle Database Cloud Service
Backup and Disaster Recovery
On Premises
Copyright © 2015, Oracle and/or its affiliates. All rights reserved. |
Oracle Database Cloud Services - Roadmap
Capability Database Cloud Services MAA Components
Zero downtime server High Availability (HA) Extreme Performance & Exadata Service Oracle RAC
Zero data loss HA & Disaster Recovery (DR) DR as a Service – included in all Enterprise Editions of Database Cloud Service
Data Guard or Active Data Guard*
Zero downtime maintenance and migration GoldenGate Cloud Service Oracle GoldenGate*
Zero data loss backup and recovery Backup Cloud Service Recovery Appliance*
7
Identical Oracle MAA Capabilities: On-Premises, Cloud, and Hybrid Cloud
* Planned
Copyright © 2015, Oracle and/or its affiliates. All rights reserved. |
Database
Oracle SaaS Applications
• Database and Application Tier
• Full Oracle MAA implementation – Clusters provide HA for recoverable outages – Database replication provides HA/DR – Automation of complete site
switchover/failover – High quality of service provided by Oracle
Engineered Systems – Oracle MAA configuration and operational
best practices across all tiers
8
Complete Oracle MAA Implementation in Production Today Internet
Load Balancer
Mid-Tier Applications
Load Balancer
Mid-Tier Applications
Database
Active Data
Guard
ZFS Replication
Primary Site Standby Site
Copyright © 2015, Oracle and/or its affiliates. All rights reserved. |
Program Agenda
Oracle MAA
MAA Best Practices
Wrap / Resources
1
2
3
9
4
Copyright © 2015, Oracle and/or its affiliates. All rights reserved. |
MAA Best Practices
• Well known, proven, and tested • Engineered into all Oracle products
–Software and Engineered Systems • Validated with every new release • Documented in MAA whitepapers and key Oracle Support (MOS) notes
Always Moving Forward
10
Copyright © 2015, Oracle and/or its affiliates. All rights reserved. |
MAA Best Practices
11
Active Data Guard
Copyright © 2015, Oracle and/or its affiliates. All rights reserved. | 12
Best Disaster Protection, Real-time Failover, High ROI
Active Data Guard
Primary Database Open Read-Write
Active Standby Database Open Read-Only
Primary Oracle
Instance
Database Files
Standby Oracle
Instance
Database Files
Oracle-aware Replication Continuous Oracle Data Validation
Copyright © 2015, Oracle and/or its affiliates. All rights reserved. |
Data Guard Monitoring Best Practices
• Oracle Enterprise Manager is the simplest way to monitor a Data Guard / Active Data Guard configuration
• But what if you don’t use OEM? – New MAA Best Practices show you how to accurately determine transport or apply
lag using either the Data Guard Broker or SQL *Plus command line – Uses v$dataguard_stats on standby – Explains how to utilize compute_time and datum_time to assess if a standby has
become disconnected – Also includes guidance for monitoring the overall health of the Data Guard
configuration
13
What’s New if using Data Guard Broker or SQL Command Line
MAA Best Practices: for Monitoring a
Data Guard Config ( embedded url)
Copyright © 2015, Oracle and/or its affiliates. All rights reserved. |
Data Guard Configuration Best Practices
• New MAA Best Practice paper for both 11.2 and 12.1 • Diagnosing and tuning ASYNC transport
– Network performance – Identify bottlenecks
What's New For ASYNC Transport
14
MAA Best Practices: Data Guard
Asynchronous Redo Transport
( embedded url)
Copyright © 2015, Oracle and/or its affiliates. All rights reserved. |
Data Guard Configuration Best Practices
• Diagnosing reasons for ASYNC transport lag – Using AWR to assess peak redo rate can be misleading due to averages bring down
the rate over longer period of time – Examine the time spent in each log to determine the peak redo rate on a finer level
What's New For ASYNC Transport
15
Copyright © 2015, Oracle and/or its affiliates. All rights reserved. |
Data Guard Configuration Best Practices
• New MAA Best Practice paper for both 11.2 and 12.1 • Deep dive on SYNC performance tuning
– Test results that illustrate performance gains when using best practices – For example, proper online log file sizes with a large banking customer improved
performance by 30% – Frequent log switches force a checkpoint on the standby which results in increased
I/O thereby affecting performance – Single member standby redo log placed on fast storage
What’s New For SYNC Transport
16
MAA Best Practices: Data Guard
Synchronous Redo Transport
( embedded url)
Copyright © 2015, Oracle and/or its affiliates. All rights reserved. |
Data Guard Configuration Best Practices
• New oratcptest tool • Assess network performance for
Data Guard transport, instantiation, RMAN backup, etc. – Can simulate async and sync transport – Vary payload type, size, and content
What’s New – Network Performance Tool
17
My Oracle Support: MOS Note 2064368.1
Measuring Network Capacity using oratcptest
( embedded url)
Copyright © 2015, Oracle and/or its affiliates. All rights reserved. |
Data Guard Configuration Best Practices
• New whitepaper covers redo apply performance for both 11.2 and 12.1 • How to tune with examples for various scenarios
– Tuning using top five wait events – Test results that illustrate performance gains when using best practices
• New installation and usage instructions for standby statspack – RAC considerations documented in the paper – Using in-memory ASH report when you can’t install standby statspack
What’s New For Redo Apply
18
MAA Best Practices: Data Guard Redo
Apply ( embedded url)
Copyright © 2015, Oracle and/or its affiliates. All rights reserved. |
Data Guard Configuration Best Practices
• New MAA Best Practices for both 11.2 and 12.1 • Discussion on how switchovers have changed for 12.1
– Single command – New validate option prior to switchover
• How to assess your role transition timings and where the time is being spent • Optimizations to make role transitions faster with 11.2.0.2 and higher
– _defer_eor_orl_arch_for_so"=TRUE – _switchover_to_standby_option"= OPEN_ALL_IGNORE_SESSIONS – Default behavior for 11.2.0.4 and 12.1
What’s New For Role Transition
19
MAA Best Practices: Data Guard
Switchover and Failover
( embedded url)
Copyright © 2015, Oracle and/or its affiliates. All rights reserved. |
Data Guard Configuration Best Practices
• New whitepaper covers outage repair for both 11.2 and 12.1
• Detailed discussion of how Data Guard handles various outages – Test results that show outage repair
times – How to perform the test and what to
expect
What’s New For Automatic Outage Repair
20
MAA Best Practices: Automatic Resolution of Outages to Restore Zero
Data Loss Protection ( embedded url)
Copyright © 2015, Oracle and/or its affiliates. All rights reserved. |
MAA Best Practices
21
Application Continuity
Copyright © 2015, Oracle and/or its affiliates. All rights reserved. |
Making Planned Maintenance Transparent
Operational Steps to Move Connections using Oracle RAC Stop services normally (not using FORCE option) on the node to be maintained
Disconnect long-running sessions after the current transaction completes (DBMS_SERVICE.DISCONNECT_SESSION)
Shutdown the database instance and perform maintenance
Start up instance(s) on the node
Start services that were stopped.
Observe that connections are appearing on the service again
Repeat this procedure for each node or instances that need to be shutdown for maintenance
Session Pool Draining - Does Not Require Application Continuity
My Oracle Support: MOS Note 1593712.1 Graceful Application
Switchover in RAC with No Application Interruption
( embedded url)
Copyright © 2015, Oracle and/or its affiliates. All rights reserved. |
OLTP Workload During RAC Planned Maintenance
0
1000
2000
3000
4000
5000
6000
21:4
0:56
21
:45:
05
21:4
9:14
21
:53:
23
21:5
7:32
22
:01:
42
22:0
5:51
22
:10:
00
22:1
4:09
22
:18:
18
22:2
2:27
22
:26:
36
22:3
0:45
22
:34:
54
22:3
9:03
22
:43:
13
22:4
7:22
22
:51:
31
22:5
5:40
22
:59:
49
23:0
3:58
23
:08:
07
23:1
2:16
23
:16:
25
23:2
0:34
23
:24:
44
23:2
8:53
23
:33:
02
23:3
7:11
23
:41:
20
23:4
5:29
23
:49:
38
23:5
3:47
23
:57:
56
0:02
:05
0:06
:15
0:10
:24
0:14
:33
0:18
:42
0:22
:51
0:27
:00
0:31
:09
0:35
:18
0:39
:27
0:43
:36
0:47
:45
0:51
:55
0:56
:04
1:00
:13
1:04
:22
1:08
:31
1:12
:40
1:16
:49
1:20
:58
1:25
:07
1:29
:16
1:33
:25
1:37
:35
Transaction Per Second
No impact to application during service stop/starts
Copyright © 2015, Oracle and/or its affiliates. All rights reserved. |
End-to-end Application Availability
• Replay in-flight work on recoverable errors
• Masks many hardware, software, network, storage errors and outages
• Improve end-user experience and productivity without requiring custom app development
• 12.1 supports Java • 12.2 adds support for OCI, unmanaged ODP.NET,
and XA when not part of global transaction
• Supports RAC, RAC One and Active Data Guard
Application Continuity for Unplanned Outage
Application Server
DB Servers or Data Centers
24
1. Database Request
5. Response
3. Errors, Timeouts
2. DB Calls 4. Application Continuity
Technical Whitepaper: Application Continuity with
Oracle Database 12c ( embedded url)
Copyright © 2015, Oracle and/or its affiliates. All rights reserved. |
With Application Continuity slight delay during instance failover - but no error
Without Application Continuity – application errors and connection closed
Application Continuity Seamless failover of application connections with RAC
Copyright © 2015, Oracle and/or its affiliates. All rights reserved. |
MAA Best Practices
26
Multitenant
Copyright © 2015, Oracle and/or its affiliates. All rights reserved. | 27
Architecture for consolidating databases and simplifying operations Oracle Multitenant
GL OE
AP Self-contained PDB for each application • Portability (via pluggability) • Rapid provisioning (via clones) • Applications run unchanged • PDB upgrades via plug/unplug
Shared memory and background processes • More applications per server
Common operations performed at CDB level • Manage many as one (upgrade, backups, HA) • Granular control when appropriate • Simple DR
PDBs
Root
CDB
MAA Best Practices: Database Consolidation with Oracle Multitenant
( embedded url)
MAA and Multitenant • Solutions for planned / unplanned outages
Copyright © 2015, Oracle and/or its affiliates. All rights reserved. |
Oracle Multitenant
• Prior to 12.1.0.2 all files associated with a new PDB had to be in place on the standby – If not, apply would stop and you would need
to copy over new files to the standby to continue
• 12.1.0.2 CREATE PLUGGABLE DATABASE statement has a new clause, STANDBYS=NONE – Allows for deferral of file instantiation on the standby – Structure of the PDB created on but files belonging to the PDB are marked as OFFLINE/RECOVER at the
standby – Affected PDB cannot be opened in this state, all other PDBs can be opened Read/Only with Active Data
Guard
What’s New - Operational Best Practices
My Oracle Support: MOS Note 1916648.1
Using Deferred PDB Recovery and STANDBYS=NONE with
Oracle Multitenant ( embedded url)
Copyright © 2015, Oracle and/or its affiliates. All rights reserved. |
Oracle Multitenant
• Data Guard considerations with Multitenant – Data Guard generally handles PDB plugin and
creation without need for user intervention
• Support note addresses operational best practices for cases where manual intervention is required – Cloning existing PDB to new PDB in same container database – Cloning a PDB from another CDB via Manifest File – Performing a remote PDB clone via Database link
What’s New – Operational Best Practices
My Oracle Support: MOS Note 2049127.1
Data Guard with Oracle Multitenant
( embedded url)
Copyright © 2015, Oracle and/or its affiliates. All rights reserved. |
MAA Best Practices
30
Database In-Memory
Copyright © 2015, Oracle and/or its affiliates. All rights reserved. |
Oracle Database In-Memory: Dual Format Database
• BOTH row and column formats for same table
• Simultaneously active and transactionally consistent
• Analytics & reporting use new in-memory column format
• OLTP uses proven row format
31
Normal Buffer Cache
New In-Memory Format
SALES SALES Row
Format Column Format
SALES
Copyright © 2015, Oracle and/or its affiliates. All rights reserved. |
Database In-Memory
• New MAA Best Practices for Database In-Memory – Configure and manage connections – Duplicate In-Memory column store across nodes
• Database In-Memory on generic hardware – Instance or node outage will result in In-Memory column store being repopulated
automatically – Queries begin using the IM column store as data is populated – Application back to full performance once population has completed
High Availability Best Practices
32
MAA Best Practices: Oracle Database
In-Memory ( embedded url)
Copyright © 2015, Oracle and/or its affiliates. All rights reserved. |
Database In-Memory
• Database In-Memory on Exadata / SuperCluster – Tables can be mirrored in-memory across instances / nodes – Data stored in-memory doesn’t need to be re-populated if instance or node goes
down – Simply specify the DUPLICATE sub-clause of the INMEMORY attribute
• Two methods to duplicate – Duplicate = mirror column store on one other node – Duplicate all = table is replicated across all nodes in the cluster
High Availability Best Practices
33
Copyright © 2015, Oracle and/or its affiliates. All rights reserved. |
Database In-Memory
• FAN managed services push HA even further – Automatically disable or stop services on nodes where
a repopulation is occurring – Services continue to run on nodes where data is in-memory – Once restarted instance has completed population services
automatically start
• Sample scripts to automate stopping / starting services
High Availability Best Practices
34
My Oracle Support: MOS Note 1927000.1
Automatically Disable a Service Based on FAN Events as Needed
( embedded url)
Copyright © 2015, Oracle and/or its affiliates. All rights reserved. |
MAA Best Practices
35
GoldenGate
Copyright © 2015, Oracle and/or its affiliates. All rights reserved. |
Flexible Logical Replication Oracle GoldenGate
• Zero-downtime maintenance and migrations • Active-Active high availability • Heterogeneous replication, data distribution and integration
Source & Target Oracle & Non-Oracle Database(s)
Target & Source Oracle & Non-Oracle Database(s) Bi-directional
LAN / WAN / Internet Over TCP/IP
36
Copyright © 2015, Oracle and/or its affiliates. All rights reserved. |
Oracle GoldenGate
• Transparent Role Transitions in a Data Guard Configuration – No manual intervention is required with FSFO and DG Broker
• Configuration makes use of: • Oracle Grid Infrastructure Bundled Agent (XAG) • DBFS for shared GoldenGate files (trails and checkpoint files) • Role based services • Integrated Extract (with HANDLEDLFAILOVER option for ASYNC DG) • Integrated Replicat is recommended
What’s New
37
MAA Best Practices: Transparent Role
Transitions with Data Guard and Oracle
GoldenGate ( embedded url)
Copyright © 2015, Oracle and/or its affiliates. All rights reserved. |
Sample Deployment
38
Observer
Primary Database Standby Database
Redo Transport (SYNC or ASYNC)
Integrated Extract LogMining
Server
Trail and other OGG Files In DBFS
Redo Transport
OCI Connection
File I/O Warehouse
Bidirectional GoldenGate Replication
Copyright © 2015, Oracle and/or its affiliates. All rights reserved. |
Sample Deployment – Post Role Transition
39
Observer
(OLD) Primary Database (NEW) Primary Database
Redo Transport (SYNC or ASYNC)
Integrated Extract LogMining
Server
Trail/Checkpoint/BR Files In DBFS
LogMining Server
Redo Transport
OCI Connection
File I/O
Warehouse
Bidirectional GoldenGate Replication
Copyright © 2015, Oracle and/or its affiliates. All rights reserved. |
MAA Best Practices
40
Oracle Cloud
Copyright © 2015, Oracle and/or its affiliates. All rights reserved. |
Disaster Recovery on the Oracle Cloud
41
Hybrid Cloud Model for Disaster Recovery
Data Guard
On-Premises Database
Standby Database
Test/Dev
Databases
Reporting
Oracle Database
Backup Service
• Instant DR site, zero capital expense • Simplifies implementing DR for
compliance & regulatory purposes • Data Guard and Active Data Guard
– Use standby for DR, reporting and testing
Copyright © 2015, Oracle and/or its affiliates. All rights reserved. |
DR on Hybrid Cloud - Today
• New MAA Best Practices – Use all existing MAA Best practices – Create security IP lists to limit access – Oracle Net encryption – Transparent Data Encryption (TDE) – Install Oracle Restart for TAF
42
MAA Best Practices: Disaster Recovery for
Hybrid Cloud - Primary on Premises, Standby in Cloud
( embedded url)
Copyright © 2015, Oracle and/or its affiliates. All rights reserved. |
Customer will be presented the option to provision a
standby database
Customer will be asked if primary database is on premises or on cloud
43
Cloud tooling will automate creation and
life-cycle management of standby databases for
both use-cases
DR on Hybrid Cloud – Future Plans
1 2 3
As Simple as 1-2-3
Copyright © 2015, Oracle and/or its affiliates. All rights reserved. |
Program Agenda
Oracle MAA
MAA Best Practices
Resources
1
2
3
44
4
Maximum Availability Architecture with 12c
©2015 LinkedIn Corporation. All Rights Reserved.
Sandip Davda linkedin.com/in/sandipdavda
Maximum Availability Architecture with 12c
About LinkedIn LinkedIn operates the world’s largest professional network on the Internet with more than 380 million members in over 200 countries and territories.
Connect the world's professionals to make them more productive and successful.
Mission
©2015 LinkedIn Corporation. All Rights Reserved.
Maximum Availability Architecture with 12c
Why Oracle MAA? Successful off-the-shelf Replication Products. Easy to adapt.
©2015 LinkedIn Corporation. All Rights Reserved. 47
Primary Use Case
High availability of LinkedIn data, with multiple sites serving Read/Write traffic.
Maximum Availability Architecture with 12c
MAA Architecture: LinkedIn
©2015 LinkedIn Corporation. All Rights Reserved. 48
GoldenGate
Active Data
Guard
Active Data
Guard R/O
R/O
R/O R/O
Maximum Availability Architecture with 12c
MAA with Golden Gate N-way Active-Active. Timestamp and priority -based CDR.
– Added columns to handle CDR. – GG columns updated by triggers, independent of Application.
Non-overlapping primary keys/unique columns, across colos.
Parallelism for scaling. Soft Deletes to avoid rebirth scenarios.
©2015 LinkedIn Corporation. All Rights Reserved. 49
Maximum Availability Architecture with 12c
MAA with Active Data Guard
Disaster Recovery RMAN Backups For Read Only Traffic
- Boot strapping cache. - Allowing QA/developers to access production data.
Sending data to down stream such as Hadoop, Teradata, etc for data analytics and reporting.
©2015 LinkedIn Corporation. All Rights Reserved. 50
Maximum Availability Architecture with 12c
12c Features at Use
©2015 LinkedIn Corporation. All Rights Reserved. 51
Encrypting data with the Master Key and Wallet Method Encrypts GG trail file while its being transmitted over TCP/IP.
Storage Snapshot Optimization Allows Third Party Snapshot technologies to take storage snapshots without putting the database or associated data files in BACKUP mode.
Recovering tables and table partitions
Very useful in logical corruption scenarios, no need to restore entire database or tablespace.
Maximum Availability Architecture with 12c
What Next?
©2015 LinkedIn Corporation. All Rights Reserved. 52
More active colos
Upgrading all the existing databases to 12c to leverage new features.
Performance improvements, improved manageability Using Integrated
Extract/Replicat. Easy to implement Conflict Resolution using Auto CDR. Auto Sharding using Goldengate and Data Guard
Maximum Availability Architecture with 12c ©2015 LinkedIn Corporation. All Rights Reserved. 53
Copyright © 2015, Oracle and/or its affiliates. All rights reserved. |
Program Agenda
Oracle MAA
MAA Best Practices
Wrap / Resources
1
2
3
54
4
Copyright © 2015, Oracle and/or its affiliates. All rights reserved. | 55
Additional Resources
www.oracle.com/goto/maa www.oracle.com/goto/ha
Copyright © 2015, Oracle and/or its affiliates. All rights reserved. |
Safe Harbor Statement The preceding 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.
56
Copyright © 2015, Oracle and/or its affiliates. All rights reserved. | 57