ha/dr for sap - ibm · pdf fileha/dr for sap designing and operating resilient ... hot standby...

Place photo here

HA/DR for SAP

Designing and operating resilient SAP systems

Insert footer information such as IBM Confidential (View>Header and Footer) 2 2

Content

SAP HA scenarios

Concepts

Strategy - SAP Application Business Continuity

Possible Failures (SPOF)

IBM Solutions for SAP HA

Q & A


Content

SAP HA scenarios

Concepts




Q & A

Insert footer information such as IBM Confidential (View>Header and Footer) 4


High Availability Concepts

•  SAP HA scenarios require at least one Switchover Solution •  With switchover, services can be automatically switched from a failed host to

a ‘standby host’ •  allowing continuation of SAP system operations •  Switchover solutions must support SAP applications, other infrastructure,

RDBMS, server, storage, and network •  IBM has several products for SAP customers that support High Availability


High Availability –Definitions

•  Availability is the capacity to function as expected. A service is considered available if it can complete its assigned task at the appropriate time

•  This is a yes-no concept: a service is available or it is not •  Planned and unplanned downtime reduce availability •  Availability is calculated as the probability of a service being available:

– Availability = 100% * reached uptime / planned uptime – (Reliability = 100% * MTBF / (MTBF + MTTR), MTBF = Mean Time Between Failures, MTTR =

Mean Time To Recover)


Disaster Recovery

•  A disaster is a failure that prevents production at an entire location for an extended period

•  For example, a disaster might have the following causes: – Power failure – Flood – Fire – Tornado – Earthquake – War

•  Enterprises need a disaster recovery site to survive, so that production can resume quickly at a separate location


HA & DR Differences

High Availablity n  Restart < 15 minutes n  Automated Takeover n  Covers failures :

n  at local site n  protects against physical

errors n  Server n  Disks n  Adapter n  Network

n  protects against fatal SW errors

n  Operating systems n  Databases n  Applications n  Services

Disaster Recovery n  Restart ~ 4 Hours n  Manual procedures involved n  Protects against

n  loss of primary site n  Covers failure of:

n  HA solution n  primary site (Infrastructure) n  logical errors n  fatal user error n  loss of complete Infrastructure

Blocks n  Caused by

n  disaster of nature n  Heavy impact on primary site

n  Disaster Recovery Plan required


Scale of High availability

Availability Downtime per week

Weekly downtime usable for (examples)

Downtime per year

Yearly downtime usable for ….

99,9999% 0,6 sec ?? 30 sec ??

99,999% 6 sec Weekly fast switchover 5 min 1 yearly restart (?)

99,99% 1 min Daily fast switchover 52 min One offline software maintenance per year

99,9% 10 min 1 Weekly restart 8 hours 45 min One offline backup per year

99% 1h 40 min Offline software maintenance 87,5 hours

90% 16h 48 min 1 Offline backup per week 36 days

Infrastructure „Five 9s“ industry approach

Today‘s best breed Eco-system approaches

Insert footer information such as IBM Confidential (View>Header and Footer) 10 10 IBM Storage Solutions for SAP

Server Cluster

Campus I

High Availability Basic


Server Cluster

Campus I

High Availability Plus


Server Cluster

SVC stretched

cluster

Campus I

High Availability Advanced


Server Cluster

SVC stretched

cluster

Campus I

High Availability Advanced & DR Basic


Campus I

Server Cluster

SVC stretched

cluster

Campus II

High Availability Advanced & DR Advanced


Downtimes - Planned and unplanned

Source: Gartner Group 12/01

About 35% of system outages are caused by lower IT layers, i.e. HW, networking, OS

è HA concepts have to consider all levels of application availability

è To decide on appropriate HA solutions, all mission critical processes need to be analyzed with respect to the impact of one of the components‘ failure

Unplanned Outages

environ. / hw20%

Application Failure

40%

Operator Error40% Planned Outages

b'up/restore10%

environ2%

hw/net/OS13%

batch proc.10%

DB + application65%

Planned -!90% of Outages!


Causes - Unplanned Downtime

Server

Operating System

Disk Mirroring & SAN Technology

Database

Business Applications

/ /dev/vg00/vol1 /home /dev/vg00/vol3 /net <any host>:<all ex. /opt /dev/vg00/vol4

Storage

Network Redundant Network Topology

Cluster

Technology

Redundancy, Cluster

Cluster,

Replication


Causes - planned downtime

High frequency

Long duration Minutes 0,5 … 2 hours 10…15 hours

Weekly

Monthly

Quarterly

Yearly

Offline backups

without split-mirror

Offline backups

with split-mirror

Kernel upgrades Profile

parameter changes

Transports

Support packages

End of daylight

saving time Release

upgrades

Database reorganizations

Low frequency

Short duration

Daily


The End-to-End view

1.  Identify all components involved from end-user to “lowest” level to be protected against failure: >> No single point of failure

2.  Identify the core components of the data center, e.g.: -  Server HW -  Storage system -  Network: SAN / LAN -  OS -  DB -  Applications

Basically one single error of a component type can be secured.

It is possible to handle subsequent errors, but this is usually not part of High Availability.

Limits of „eco“ system


Roles of SAP, IBM, and Customer

•  SAP –  Provides HA capable integration and application platform (SAP NetWeaver)

and application scenarios on top of SAP NetWeaver •  Leveraging the HA capable computing infrastructure •  Working with partners regarding platform specific procedures •  provide MSCS specific HA procedures

•  IBM –  Provide HA capable computing infrastructure

•  Hardware, operating system, database system, file system … –  Provide platform specific HA procedures

•  Configuration, switch-over •  Implementation at customer site

•  Customer –  Define required HA levels –  Provide IT management concept and guidelines –  Ensure appropriate operating procedures, training of IT staff

à  High availability is a joint responsibility of SAP, IBM, and customer


Content

SAP HA scenarios

Concepts

Strategy - SAP Applications Business Continuity



Q & A


SAP infrastructure - Resiliency

CI

repl.enqueue

Replicated enqueue cluster.

DB

Cluster or standby DB

DB reconnect

DB

RDBMS

CIMessageServerEnqueueServer

WebDisp.

Transparent load

balancing

WebDisp.

(DMZ)

Appl.Server

Appl.Server

HA clusters

ABAP J2EE

ABAP J2EE

ABAP J2EE

RDBMS

RDBMS RDBMS

Portal / PI

D/R

SCS SCS‘


SAP‘s HA Strategy

–  SAP‘s HA scope: •  business applications – Business continuity •  technology components

–  Infrastructure components (network – server – storage – DB) •  partner products with their HA features

–  Unplanned downtime •  eliminate SPOFs

–  Planned downtime •  to be decreased with smart software logistics

and rolling maintenance procedures

Storage

Network Server

OS Data Base

Part

ner S

olut

ions

NetWeaver

SAP Business Suite


SAPGUI

R/3 Application Servers F Message Server F Enqueue Server

Database Server

Traditional SAP R/3 High Availability Requirements

Network / Backbone

R/3 Data

IBM

Few, well defined number of mission critical components in classical R/3 landscape!

SAPGUI

IBM

IBM

IBM


SAPGUI

Application Servers

Database Server

IBM High Availability solutions whole SAP landscape

Network / Backbone

SAP Data

Functional Backends F parallel DBs F liveCache F etc.

F Enterprise Portal F PI F ITS/IGS

Transient Functions

ABAP-Engine F Message Server F Enqueue Server

JAVA-Engine F Message Server F Enqueue Server

Replicated Enqueue Server

Hot-Standby liveCache

DB2 HADR DB2 PureScale

Oracle RAC

PowerHA

TSA


High Availability Clustering Solutions •  High Availability (HA) Management

–  Business continuity by workload failovers –  Provides for redundant operating environment

–  Automated workload bring ups –  Environment specific outage actions

–  Planned and unplanned outage management

•  Disaster Recovery (DR) Management –  Geographically dispersed locations

–  Host based or Storage based replication –  Simplified compliance testing

Data Center Clustering

External LAN

LAN 1 LAN 2

SAN 1

SAN 2

Heartbeat 1

Heartbeat 2

Logical server

Multi-Site Clustering

External LAN

LAN 1 LAN 2

SAN 1

SAN 2

Heartbeat 1

Heartbeat 2

Logical server

External LAN

LAN 1 LAN 2

SAN 1

SAN 2

Heartbeat 1

Heartbeat 2

Logical server


Content

SAP HA scenarios

Concepts


Possible Failures(SPOF)


Q & A


•  Several studies place the proportion between 20% and 45% •  Human error, software error and planned maintenance cause the majority of service outages

Hardware Failures Account for a Small Minority of System Outages

SAP Single Point of failures (SPOF) •  SAP Central Services - Message/enqueue server •  Global file services - /sapmnt, /usr/sap/trans •  Database •  NFS Server •  Optional components:

–  SAP Web Dispatcher –  SAProuter

• Network failure –  Outage of a Network Interface Card

• Disk/Storage failure –  Disk outage

• Hardware failure –  System outage

Failure scenarios


SAP application features to minimize downtime

•  Logon groups decouple physical servers from User front-end •  DB reconnect of App-Servers •  Virtualization of hostnames •  Replicated Enqueue for faster SCS take-over

•  Transaction & Data integrity on application side •  Rolling kernel upgrade

– for SCS maintenance •  Enhancement Packages Policy

– New Business Logics and Patches

– Switch Framework allows for selection

Network

Storage Server

OS Database

Part

ner S

olut

ions

NetWeaver

SAP Business Suite


Content

SAP HA scenarios

Concepts




Q & A


–  IBM HACMP (AIX) –  IBM HACMP/XD (AIX) –  IBM System Automation (OS/390) –  IBM Tivoli System Automation for Multi Platform –  IBM iASP (OS/400, Release V5R2 and newer) –  Fujitsu Siemens: PRIMECLUSTER, ReliantCluster (Reliant, Solaris,

Linux) –  Microsoft Cluster Service (Windows) –  SteelEye LifeKeeper (Linux) –  Veritas Cluster Server (several OS)

Cluster solutions on IBM Platform

–  IBM PowerHA –  IBM System Automation (OS/390) –  IBM Tivoli System Automation for Multi Platform –  IBM iASP (OS/400, Release V5R2 and newer) –  Fujitsu Siemens: PRIMECLUSTER, ReliantCluster (Reliant, Solaris,

Linux) –  Microsoft Cluster Service (Windows) –  SteelEye LifeKeeper (Linux) –  Veritas Cluster Server (several OS)


DB2 HADR Asynchronous,log-based replication of a database

Replicated standby db for DB2 for zOS Synchronous & asynchronous replication of database

Oracle Dateguard Asynchronous log-based replication of a db to one site

Symmetric replication for oracle -asynchronous and synchronous statement-based replication of data to one or

more sites

Microsoft SQL server standby database Asynchronous,log-based replication of a database

Hot Standby for MaxDB/SAPDB - Near synchronous,use storage device for log sharing

Replication Technologies

IBM Metro Mirror

IBM Global Mirror

IBM Global Copy

IBM Metro Cluster

Compaq DRM

EMC SRDF

Hitachi HRC

HP Continuous Access XP Sun StorEdge Network Data Replication

Veritas Volume Replicator

DB Level - Products Storage Level - Products


Definition of PowerHA Power Systems High Availability Solution For mission critical application availability through planned and unplanned outage events Shared Storage Clustering Technology designed for automation and minimal IT operations. Embedded technology for integrated simplicity and reliability


Technically - PowerHA •  Deep Integration –  Cluster Aware AIX

–  Kernel based event processing

–  Centralize cluster topology management

•  Ease of deployment & ease of use –  Systems Director based management

–  Discovery based deployment

–  Cluster wide security

•  Multi-Site & Disaster Recovery –  Differentiate with IBM storage

–  Heterogeneous storage support

•  Solution Package Optimization –  Standard Edition, Enterprise Editions


Advantages of PowerHA Simpler to deploy and easier to manage with IBM Systems Director, intuitive interfaces, cluster and resource group wizards, management dashboards and Smart Assists for SAP and other popular applications Minimize IT operations with cluster aware AIX; cluster wide AIX commands, kernel based event management, device naming, central repository and multi-channel communications Robust cluster integrity with disk fencing and multi-channel heart beat which automatically uses available I/O including SAN. Complete end to end failover automation with policy driven resource group relationship sequencing


ü  Minimizes planned and unplanned outage events –  Automated for action and simplicity

ü  Simplifies cluster administration –  Embedded cluster technology –  Graphical user interface

ü  Designed for both data center and multi-site deployments –  Manages the most prevalent storage solutions –  Options for stretch clusters and linked clusters

ü  Minimizes cost factors –  DLPAR manages processor entitlement licensing for

cost optimization between primary and secondary nodes

–  Smart Assists for application configuration and management included at no charge

•  Compliance testing automation to minimize operations impact

Customer situations

Ø  Stringent service level agreements, no tolerance for downtime

Ø  OS & HA management complexity

Ø  Multi-site solution management

Ø  Cost factors

PowerHA Solution Client Pain Points


NEW ! PowerHA SystemMirror 7.1 Enterprise Edition Simpler to deploy and easier to manage multi-site configurations with IBM Systems Director, intuitive interfaces, multi-site install wizard Stretched Cluster; Cluster wide AIX commands, kernel based event management single repository multicast communications Linked Clustering; cluster wide AIX commands, kernel based event management, linked clusters with unicast communications & dual repositories HyperSwap for continuously available storage in two-site topologies Cluster Split/Merge technology for managing split-site policy scenarios

Announce Date: Oct 3 2012 GA Date: Nov 16 2012


PowerHA SystemMirror 6.1 Verses 7.1

PowerHA SystemMirror 6.1 7.1 PowerHA 7.1 Benefit IBM Director based graphical user interface NA ü  Ease of Use

Cluster Aware AIX (CAA) RSCT ü  Reliability

Triple redundant heartbeat (multicast) NA ü  Effectively eliminates partitioning

SAN based communications NA ü  Additional cluster communication path

Stretched cluster (shared repository) NA ü  Two-Site multicast HA/DR shared network

Cross Site Mirroring (single site stretch cluster) NA ü  LVM mirroring with CAA

Linked clusters (separate repositories) NA ü  Two-Site HA/DR separate networks

HyperSwap with DS8800 NA ü  Two-Site continuously available storage

Multi-Site set up wizard NA ü  Speeds up implementation

Tie Breaker (SOD) NA ü  Can eliminate split-site scenarios

Federated Security NA ü  Cluster wide security management

Live Cache SAP hot standby NA ü  Fast failover for APO SCM


Sync mode Async mode Active Active Consistency groups

GLVM Yes Yes ü  ü 

XIV Metro Mirror Global Mirror ü  ü 

Storwize® V7000 Metro Mirror Global Mirror ü  ü 

SVC Metro Mirror Global Mirror ü  ü 

DS8800 Metro Mirror Global Mirror ü  ü 

EMC SRDF SRDF ü  ü 

Hitachi TrueCopy Universal Replicator

TrueCopy Universal Replicator

ü  ü 

HP Storage Works P9500

Truecopy Universal Replicator

Truecopy Universal Replicator

ü  ü 

HP XP12000, XP24000 Continuous Access Software replication

Continuous Access Software replication

ü  ü 

•  Solutions for disaster recovery –  Major vendors, popular options –  IBM storage options that span the market

•  Statement of Direction (SOD)

PowerHA SystemMirror 7.1 Enterprise Edition Storage Options for your Enterprise

SOD

IBM intends to update IBM PowerHA SystemMirror Enterprise Edition v7 to support: select EMC storage subsystems that provide Symmetrix Remote Data Facility (SRDF) replication services select Hitachi storage subsystems that provide TrueCopy or Universal replication services select HP storage subsystems that provide Continuous Access Software replication services


SAP Solution Package IBM PowerHA SystemMirror 7.1 •  Motivation

–  Standardized toolset around the globe –  Support for replicated enqueue scenarios –  Align implementation to SAP recommendations –  Provide SAP specific Best Practices and recommendations

•  Supported SAP scenarios –  ABAP, JAVA and Double Stack w/o ERS and App (optional)

•  Tested for NW7.0, 7.20, ECC6.0, EP6, PI7.1 –  2-tier and 3-tier installations –  Multi-node clusters –  DB2 and Oracle DBs –  HA + DR (PowerHA and SystemMirror Enterprise)

•  Solution –  Set of documentation covering Storage, VIOS, AIX and PowerHA for SAP HA installation & configuration –  PowerHA configuration –  PowerHA start, stop and monitor scripts –  Distributed for free to IBMers and BPs via request to ISICC Infoservice

•  Limitation –  manual setup (not automated) of PowerHA accordingly to Documentation

Automation with PowerHA

SystemMirror 7.1


PowerHA SystemMirror Smart Assist

Smart Assist for DB2 DB2 UDB Enterprise server edition - 8.1,8.2,9.1 or 9.5

Smart Assist for Oracle Oracle Application server 10g Rel 2, 11g Release 2

Smart Assist for MaxDB -  SAP MaxDB 7.6 and 7.7

-  SAP Live Cache - SAP MAXDB version 7.7.07 or later DSCLi for DS8000

Smart Assist for SAP –  SAP NW- 7.0, 7.1,7.2 and 7.3 verison

Smart Assist features: ► Wizard for selected setups ► Discovery of running instances ► Automated cluster configuration incl.scripts ► IBM Systems Directors integrated

• PowerHA SystemMirror 7.1 or later versions • AIX - Version 6.1 with 6100-06 Technology level or later • Available Smart Assist for ( DB2,Oracle,MaxDB) • Smart Assist for SAP

What is Smart Assist ?


Smart Assist for SAP Infrastructure

•  Discovers SAP Deployment in the Cluster –  Varied deployments whether 2 or 3 tier are covered

•  Guides customer with HA policy creation –  Out of the box configuration policies

–  Methods to manage the start/stop of the environment –  Health monitoring methods

–  Dependency and relationship discovery on other OS resources •  Support for:

–  DB2, Oracle or MAXDB databases –  Supports policy changes for customized deployments

–  Supports both ABAP and Java Infrastructure

Typical SAP Deployment involves: ► Multiple Compute Servers ► Database instances ► Shared File Systems ► Two tier (Application services and Database reside on the same OS Image) and three tier deployments (Application server and database reside on different OS images)

NFS

PowerHA Cluster

/usr/sap/trans

/sapmnt/<SID>

Primary Node:CI InstanceDialog Instance

Stand by Node:

Enqueue Server

Message Server

ABAP SCS Instance

NFS

Virtual IP for SCS

Virtual IP

Central/Dialog Instance

Virtual IP

Central/Dialog Instance

Enqueue Server

Message Server

Java SCS Instance

Resource Group for SCS Instance

Shared Disk

/usr/sap/<SID>/ASCS<nn>

/usr/sap/<SID>/SCS<nn>

Shared Disk

/usr/sap/<SID>/D<nn>

Resource Group for Central/Dialog Instance

Enqueue Server

Message Server

Virtual IP for SCS

Enqueue Server

Message Server

ABAP SCS Instance Java SCS Instance

Note: ERS is not shown in the figure but will be handled by HACMP with an another resource group.

Shared Disk

/export/sapmnt/<SID>/export/usr/sap/trans>

Resource Group for SAP Global FS

Virtual IP for NFS

• PowerHA integrated solution: Optimized management interface in SMIT and Director

• Support for 2 or more node SAP clusters • Exploitation of database smart assists, exploits advanced RG policies • Robust handling of ERS instance in a 3/more node cluster


DB2 HADR – Basic Principle

Standby

DB2 Engine

HADR HADR

log writer

Primary

DB2 DB2

log writer

DB2 Engine

•  Two active machines Ø  Primary

•  Processes transactions •  Ships log entries to Standby

Ø  Standby •  Cloned from the primary •  Receives and stores log entries from the

primary •  Re-applies the transactions

•  If the primary fails, the standby can take over the transactional workload

•  If the failed machine becomes available again, it can be resynchronized


IBM Tivoli System Automation for Multiplatforms provides policy-based application and resource self-healing

Manages outage of landscapes for business application availability –  HW/OS/SW independence important for SOA/ESOA

landscapes –  Fast detection of outage through monitoring –  Knowledge about the critical system landscape,

which is managed –  Quick and consistent recovery of failed resources

and whole applications either in place or on another system

–  Policy allows flexible adaptation to customer individual BC necessities which are different customer by customer

TSA


DB2 HADR & TSA a Solid Combination

Scenarios: •  Onsite Standby •  Offsite Disaster Recovery Features: •  SAP and DB2 HA & DR •  Long Distance DR •  Failover times in seconds •  If the primary fails, the secondary takes

over the transactional workload •  If the failed machine becomes available

again, it can be resynchronized •  Rolling upgrades •  Automatic client reroute

Classic Active / “Passive” Fail-Over Cluster

DB

Ser

ver

AS

AS (CI) (DB)

Boot

Shared Disk

Boot

AS AS

CI (ENQ) ENQ

CI

AS (DB)

DB Server Secondary D

B S

erve

r Pr

imar

y

AS

Boot

AS AS

Disk Disk

New HADR Fail-Over Cluster

Boot

Replication

(CI)


HADR & TSA Integration - Local Implementation


DS48E01

FC32E01

NWE01

DS48E02

FC32E02

NWE02

DB Replication withHACMP / HADR

Application Failover withHACMP / Repl. enqueue

SMP01

VIR

TU

AL

IO

La

ye

rV

IOS

VIO

S

Sandbox

Dev. SystemsLocal Standby

Test SystemsActive

ME

MO

RY

AN

D C

PU

RE

SO

UR

CE

S (

8 x

P5

1.6

5 +

96

GB

Me

mo

ry)

HY

PE

R V

ISO

R D

YN

AM

ICA

LL

Y D

IST

RIB

UT

ING

RE

SO

UR

CE

S

P570E01

PH

YS

ICA

L IO

LA

YE

R (

18

x F

C +

12

x N

IC

BWP01

HRP01

MEP01

ME

MO

RY

AN

D C

PU

RE

SO

UR

CE

S (8

x P

5 1

.65

+ 9

6 G

B M

em

ory

)

HY

PE

R V

ISO

R D

YN

AM

ICA

LL

Y D

IST

RIB

UT

ING

RE

SO

UR

CE

S

P570E02

PH

YS

ICA

L IO

LA

YE

R (1

8 x

FC

+ 1

2 x

NIC

AUP03

SMP02

VIR

TU

AL

IO

La

ye

rV

IOS

VIO

S

HRP02

MEP02

Dev. SystemsGlobal Standby

Sandbox

Dev. SystemsActive

Test SystemsLocal Standby

Test SystemsGlobal Standby

DS48U01

FC32U01

NWU01

DS48U02

FC32U02

NWU02

DB Replication withHACMP / HADR

Application Failover withHACMP / Repl. enqueue

ME

MO

RY

AN

D C

PU

RE

SO

UR

CE

S (8

x P

5 1

.65

+ 9

6 G

B M

em

ory

)

HY

PE

R V

ISO

R D

YN

AM

ICA

LL

Y D

IST

RIB

UT

ING

RE

SO

UR

CE

S

P570U02

PH

YS

ICA

L IO

LA

YE

R (1

8 x

FC

+ 1

2 x

NIC

VIR

TU

AL

IO

La

ye

rV

IOS

VIO

S

AUP02

BWP02

MEP03

Sandbox

Dev. SystemsLocal Standby

Test SystemsActive

Dev. SystemsGlobal Standby

SMP03

ME

MO

RY

AN

D C

PU

RE

SO

UR

CE

S (

8 x

P5

1.6

5 +

96

GB

Me

mo

ry)

HY

PE

R V

ISO

R D

YN

AM

ICA

LL

Y D

IST

RIB

UT

ING

RE

SO

UR

CE

S

P570U01

PH

YS

ICA

L IO

LA

YE

R (

18

x F

C +

12

x N

IC

HRP03

VIR

TU

AL

IO

La

ye

rV

IOS

VIO

S

AUP01

BWP01

Sandbox

Dev. SystemsActive

Test SystemsLocal Standby

Test SystemsGlobal Standby

Log Shipping

TSA/HADR TSA/HADR

TSA/Rep. enqueue TSA/Rep. enqueue

HADR & TSA Integration - Global Implementation


DB2 High Availability Features

Planned Downtime •  Online Space Management •  Online Reorganization •  Online Configuration

Changes •  Online Backup •  Online DB Split/Copy •  Online Rolling Upgrade/

Updates

Unplanned Downtime •  Automatic Recovery •  Traditional Active/“Passive“

Fail-Over Clusters •  Shadow Database •  High Availability Disaster

Recovery (HADR) – Fast Fail-Over


IBM TSA cluster software

Clients

Application Servers

Primary Standby

HADR

Database and Logs

Central Instance

TSA

Primary Standby

HADR

TSA

Database and Logs

Database and Logs

Database and Logs

Texas 1 Texas 2

DB2 HADR + TSA + SAP + AIX


Tivoli System Automation for Multi Platforms

•  High Availability (HA) solution for platforms running Linux and AIX

•  System Automation guarantees high availability for business applications –  Can be used for applications of any type (Databases, WebServer, SAP, …)‏ –  Provides fast detection of HW failures and SW failures (“Monitoring”)‏ –  Performs automated recovery, like restart in place of failover (“Automation”)‏

•  Policy-based HA solution with powerful policy elements –  Allows to describe automation behavior on a high abstraction level

•  Example: Almost no effort to extend a two node scenario to eight nodes •  No script programming

Resource Mgrs

Event Coordinated Restart & Failover

Automation Mgr

Restart & Failover Rules

SAP Policy

DB2 Policy

Apache Policy

Customer Policies

Pre-Canned Scenarios: out-of-the-box policies and features provided

by System Automation

Generic Scenario: customer specified policies

…


Tivoli System Automation 3.2 Product Family

IBM Tivoli System Automation for Multiplatforms

Automation Adapter

IBM Tivoli System Automation Application Manager

Automation Adapter

Available Adapter IBM HACMP MSCS Veritas Cluster Services

Linux on System x Linux on System p Linux on System z AIX Windows Server 2008 z/OS

IBM Tivoli System Automation

for z/OS

Automation Adapter

z/OS z/OS Linux AIX Windows Server 2008 SUN Solaris

Other Clustering/HA Products: •  IBM HACMP •  MSCS •  Veritas

Linux on System x Linux on System p Linux on System z AIX Windows Server 2008 SUN Solaris

Integrated with DB2 LUW 9.5+ and installed with SAPINST NW 7.0+


AIX LVM Mirroring

•  LVM Mirroring is done by the Logical Volume Manager on the pSeries server –  The server is „aware“ that a second data copy exists. –  Both volumes have read/write access. –  A volume Failure does not cause application crash, due to I/O is redirected to the remaining data copy. –  Server can read from both volumes.

•  LVM Mirroring will also affect write performance, BUT not as much as remote copy

–  Writes to both volumes are issued in parallel. I/O is complete when the „slower“ path acknowledges that data is stored.

•  LVM Mirroring increases read performance –  LVM can balance reads across both data copies and achieve a higher degree of parallelism

•  LVM overhead is negligible, i.e. about 2% cpu load


Replication on Disk Level

Production Site

Replication Site

Host Request

Write confirm

Metro Mirror Global Mirror Global Copy Allows synchronous mirroring within a 10km (max 100km) distance between the production and replicated system

Asynchronous mirroring. No acknowledgement to host, if replicated site in error. Primary controller manages error + peak conditions (buffers Writes)

Manually triggered physical replication of flash copy è remote volume copy

Distance: near Latency: medium concurrency: 100% HA solution è phys

Distance: wide Latency: low concurrency:<100% DR solution è phys

Distance: wide Latency: low concurrency:<<100% DR solution è logical

Distance


Metro Mirror

•  Allows synchronous mirroring within a 10km (max 100km) distance between the primary and secondary system

•  Primary benefits: –  Offers one of the highest forms of protection with both primary and secondary volumes kept current

to the last update –  Note: performing the remote write may cause performance degradation


Global Mirror

•  Write operations to the secondary subsystem matches I/O completion order on the local subsystem for all volumes in the consistency group (one group supported) –  Option for Global Mirror (asynchronous write mode)

•  Primary benefits –  Maintains data integrity in multi-LUN applications (databases)

•  eliminates out-of-order updates at the secondary side

–  Note: this mode will have a performance hit compared to asynchronous write (Global Copy)


•  Allows the primary subsystem to acknowledge a host write request before the data has been successfully mirrored

Global Copy

•  Primary benefits –  Reduces impact of latency when replicating over longer distances

•  Provides performance improvement – compared to synchronous – for primary site I/O (subsystem and application)

•  Enables effective replication over longer distances (WAN)

–  Note: may result in remote site without the “latest-greatest” data


" Single identical VMs running in lockstep on separate hosts

" Zero downtime, zero data loss failover for all virtual machines in case of hardware failures

" Zero downtime, zero data loss

" No complex clustering or specialized hardware required

" Single common mechanism for all applications and OS-es

" Currently only 1vCPU. Good for CI services.

VMware vSphere™

OS SAP

OS SAP

OS SAP

VMware Fault Tolerance for SAP Solutions


Linux / Windows: Symantec Veritas HA / DR solutions for SAP

•  Veritas Cluster Server –  „Out of the box“ solution, “Push button” recovery

–  Centralized Management “single pane of glass” for

fully automatic failover and multi-site control of 100’s of clusters

–  Preconfigured agents for SAP Netweaver, Databases, mult-tier and multi-SID support

–  Heterogenous support (AIX, HP-UX, Linux, Solaris, VMware, Windows)

– Multiple configurations: •  Local Clustering •  Metropolitan HA

(Stretch Cluster) – Synchronous Replication or Mirroring

•  Wide-Area Disaster Recovery (Global Cluster) – Asynchronous Replication

-  HA and DR failover testing without affecting the primary environment:

•  Virtual Fire Drill: validates configuration via agent entry point to insure failover as expected

•  Physical Fire Drill: creates snapshot and executes failover without interrupting production or replication to DR site


Linux / Windows: Symantec Veritas HA / DR solutions for SAP

§  Storage Foundation for Windows

–  Extends Windows native logical disk management (LDM) by online storage management and dynamic disk groups

§  Storage Foundation for Windows HA –  Enhanced by Veritas Cluster Server

§  Veritas Cluster Server §  Storage Foundation HA

–  Enhanced by Veritas File System and Veritas Volume Manager

– Online storage management – Multi-pathing I/O between many server

nodes and shared storage –  In failover case, more stable storage protection by

SCSI 3 reservation

§  Storage Foundation Cluster File System –  Further enhancement by Veritas Cluster

File System –  Supports a single file system schema for

up to 32 servers –  Speeds up failovers and overcomes

NFS cross mount problems


Linux Cluster Management Software Common Features: •  Toolkit / Consulting •  Application Interface Specification (AIS) & Open Cluster Framework (OCF): Resources, Services,Dependencies, Cluster Resource Manager •  Avoid Split-Brain, Quorum, Fencing •  Host-based mirroring, MD (Multiple Device Driver, Software RAID)

As of SLES 9 •  Heartbeat 1 / 2 Open Source •  Fencing: Power (STONITH), SAN (SFEX), Ping Nodes, Failcount checks •  DRDB 8, Distributed Replicated Block Device – replication software •  EVMS, Enterprise Volume Manager

New in SLES 11 •  Billable HA extension •  Open AIS (communication) + Pacemaker (cluster) •  cLVM, cluster aware Logical Volume Manager

As of RHEL 3 •  Cluster Support Open Source •  Fencing: Power, SAN, SCSI New in RHEL 5 •  Subscription for „RHEL Advanced Platform“ (Virtualization, Red Hat Cluster Suite) •  Open AIS •  Global File System (GFS) •  CLVM, Clustered Logical Volume Mgr •  cluster aware Storage Mirroring Planned •  SAP support of Red Hat Enterprise Virtualization (Hypervisor + Monitor)


Virtualization HA Hypervisor

Cluster Support for Guests

VMware HA vSphere 4/5 RHEL 5/6

VMware HA vSphere 4/5 SLES 9 – 11

SLES 10.3+: Heartbeat 2 SLES 11+: Pacemaker

SLES 11.1+: Pacemaker

RHEL 5.3 + Cluster Suite

RHEL 5.4 + Cluster Suite

Guest Migration

VMware VMotion

VMware VMotion

Live Migration (Xen motion)

Live Migration

Live Migration (Xen motion)

Live Migration


Content

SAP HA scenarios

Concepts


Possible Failovers (SPOF)


Q & A


SAP Main Products and Solutions

65

Mail contact: [email protected]

Time for questions …


Content

SAP - HA scenarios

Concepts




Q & A


MSG ENQ

Web Application Servers

dispatcher

GW

icman

J2EE ABAP

REPL

VIPA

connect

replicate

Central Services Central Services

Scenario 1: SAP Central Services startup Policy Rule - ES, MS, VIPA collocated

Policy Rules - ERS starts after ES - ERS is anti-collocated to ES


MSG ENQ


dispatcher

GW

icman

J2EE ABAP

REPL

VIPA

connect

replicate


Scenario 2: SAP Central Services failure


MSG ENQ REPL

dispatcher

GW

icman

J2EE ABAP

VIPA

reconnect

IP takeover

Central Services Central Services rebuild


Scenario 3: Moving SAP Central Services Policy Rule - ES collocated to ERS if ERS not offline


MSG ENQ

dispatcher

GW

icman

J2EE ABAP

VIPA


REPL replicate

restart


Scenario 4: Restart Replication server Policy Rules - ERS starts after ES - ERS is anti-collocated to ES Executed if failing node is back or if a 3rd node is available


Information in this document concerning non-IBM products was obtained from the suppliers of these products or other public sources. Questions on the capabilities of non-IBM products should be addressed to the suppliers of those products. IBM may have patents or pending patent applications covering subject matter in this document. The furnishing of this document does not give you any license to these patents. Send license inquires, in writing, to IBM Director of Licensing, IBM Corporation, New Castle Drive, Armonk, NY 10504-1785 USA. All statements regarding IBM future direction and intent are subject to change or withdrawal without notice, and represent goals and objectives only. The information contained in this document has not been submitted to any formal IBM test and is provided "AS IS" with no warranties or guarantees either expressed or implied. All examples cited or described in this document are presented as illustrations of the manner in which some IBM products can be used and the results that may be achieved. Actual environmental costs and performance characteristics will vary depending on individual client configurations and conditions. IBM Global Financing offerings are provided through IBM Credit Corporation in the United States and other IBM subsidiaries and divisions worldwide to qualified commercial and government clients. Rates are based on a client's credit rating, financing terms, offering type, equipment type and options, and may vary by country. Other restrictions may apply. Rates and offerings are subject to change, extension or withdrawal without notice. IBM is not responsible for printing errors in this document that result in pricing or information inaccuracies. All prices shown are IBM's United States suggested list prices and are subject to change without notice; reseller prices may vary. IBM hardware products are manufactured from new parts, or new and serviceable used parts. Regardless, our warranty terms apply. Many of the features described in this document are operating system dependent and may not be available on Linux. For more information, please check: http://www.ibm.com/systems/p/software/whitepapers/linux_overview.html Any performance data contained in this document was determined in a controlled environment. Actual results may vary significantly and are dependent on many factors including system hardware configuration and software design and configuration. Some measurements quoted in this document may have been made on development-level systems. There is no guarantee these measurements will be the same on generally-available systems. Some measurements quoted in this document may have been estimated through extrapolation. Users of this document should verify the applicable data for their specific environment.

Revised January 19, 2006

Special Notices


Revised June 15, 2006

The following terms are registered trademarks of International Business Machines Corporation in the United States and/or other countries: AIX, AIX/L, AIX/L(logo), alphaWorks, AS/400, BladeCenter, Blue Gene, Blue Lightning, C Set++, CICS, CICS/6000, ClusterProven, CT/2, DataHub, DataJoiner, DB2, DEEP BLUE, developerWorks, DirectTalk, Domino, DYNIX, DYNIX/ptx, e business(logo), e(logo)business, e(logo)server, Enterprise Storage Server, ESCON, FlashCopy, GDDM, i5/OS, IBM, IBM(logo), ibm.com, IBM Business Partner (logo), Informix, IntelliStation, IQ-Link, LANStreamer, LoadLeveler, Lotus, Lotus Notes, Lotusphere, Magstar, MediaStreamer, Micro Channel, MQSeries, Net.Data, Netfinity, NetView, Network Station, Notes, NUMA-Q, Operating System/2, Operating System/400, OS/2, OS/390, OS/400, Parallel Sysplex, PartnerLink, PartnerWorld, Passport Advantage, POWERparallel, Power PC 603, Power PC 604, PowerPC, PowerPC(logo), Predictive Failure Analysis, pSeries, PTX, ptx/ADMIN, RETAIN, RISC System/6000, RS/6000, RT Personal Computer, S/390, Scalable POWERparallel Systems, SecureWay, Sequent, ServerProven, SpaceBall, System/390, The Engines of e-business, THINK, Tivoli, Tivoli(logo), Tivoli Management Environment, Tivoli Ready(logo), TME, TotalStorage, TURBOWAYS, VisualAge, WebSphere, xSeries, z/OS, zSeries. The following terms are trademarks of International Business Machines Corporation in the United States and/or other countries: Advanced Micro-Partitioning, AIX 5L, AIX PVMe, AS/400e, Chipkill, Chiphopper, Cloudscape, DB2 OLAP Server, DB2 Universal Database, DFDSM, DFSORT, DS4000, DS6000, DS8000, e-business(logo), e-business on demand, eServer, Express Middleware, Express Portfolio, Express Servers, Express Servers and Storage, GigaProcessor, HACMP, HACMP/6000, IBM TotalStorage Proven, IBMLink, IMS, Intelligent Miner, iSeries, Micro-Partitioning, NUMACenter, On Demand Business logo, OpenPower, POWER, Power Architecture, Power Everywhere, Power Family, Power PC, PowerPC Architecture, PowerPC 603, PowerPC 603e, PowerPC 604, PowerPC 750, POWER2, POWER2 Architecture, POWER3, POWER4, POWER4+, POWER5, POWER5+, POWER6, POWER6+, Redbooks, Sequent (logo), SequentLINK, Server Advantage, ServeRAID, Service Director, SmoothStart, SP, System i, System i5, System p, System p5, System Storage, System z, System z9, S/390 Parallel Enterprise Server, Tivoli Enterprise, TME 10, TotalStorage Proven, Ultramedia, VideoCharger, Virtualization Engine, Visualization Data Explorer, X-Architecture, z/Architecture, z/9. A full list of U.S. trademarks owned by IBM may be found at: http://www.ibm.com/legal/copytrade.shtml.

A full list of trademarks owned by SAP may be found at: http://www.sap.com/company/legal/copyright/trademark.epx UNIX is a registered trademark in the United States, other countries or both. Linux is a registered trademark of Linus Torvalds in the United States, other countries or both. Microsoft, Windows, Windows NT and the Windows logo are trademarks of Microsoft Corporation in the United States, other countries or both. Intel, Intel Xeon, Itanium and Pentium are trademarks or registered trademarks of Intel Corporation in the United States and/or other countries. AMD Opteron is a trademark of Advanced Micro Devices, Inc. Java and all Java-based trademarks and logos are trademarks of Sun Microsystems, Inc. in the United States and/or other countries. TPC-C and TPC-H are trademarks of the Transaction Performance Processing Council (TPPC). SPECint, SPECfp, SPECjbb, SPECweb, SPECjAppServer, SPEC OMP, SPECviewperf, SPECapc, SPEChpc, SPECjvm, SPECmail, SPECimap and SPECsfs are trademarks of the Standard Performance Evaluation Corp (SPEC). NetBench is a registered trademark of Ziff Davis Media in the United States, other countries or both. AltiVec is a trademark of Freescale Semiconductor, Inc. Other company, product and service names may be trademarks or service marks of others.

Special Notices (Cont.)

ha/dr for sap - ibm · pdf fileha/dr for sap designing and operating resilient ... hot standby...

Documents