SAP HANA On Power-Planning - 3.0

SAP HANA on
IBM Power Systems

and IBM System Storage
Planning Guide
IBM Deutschland Research & Development GmbH

Created on July 8, 2014 Version 0.0
Last modified on January 20st, 2016 Version 2.4
Copyright IBM Corporation, 2016
2
Note: Before using this information and the product it supports, be sure to read the general
information under Copyrights and Trademarks on page 41 as well as Disclaimer and Special
Notices on page 42.
Edition Notice and Version Information

Copyright IBM Corporation 2015. All Rights Reserved.
US Government Users Restricted Rights Use, duplication or disclosure restricted by GSA ADP
Schedule Contract with IBM Corp. All trademarks or registered trademarks mentioned herein are the
property of their respective holders.
IBM Corporation
Systems and Technology Group
Route 100
Somers, NY 10589
Produced in the United States of America
Edition Notice: 2016

This is version 2.4 of this document.
Focus: SAP HANA Scale-up and Scale-out solutions
Target: SAP HANA platform Edition SPS11 and the Power Version of SAP HANA based on SPS9 and
SPS10
Doc Version Changes

1.2 1.8 Changes documented up to version 2.1
(Jan/27/2015)
1.9 VIOS CPU sizing for production, min 4 LUNs per XFS FS, TCO optimized planning
(Aug/26/2015) prod vs. none prod, SA MP for HANA GAed
2.0 New floor for production, Introduce TDI Phase 4, relaxed non-prod server
(Sep/10/2015) prerequisites
2.1 Scale-out, SLES 11 SP4 Support
(Oct/01/2015
2.2 New SAP Notes referenced (multiple LPAR support etc.), Scale-out info, initial
(Nov/20/2015 Version for SPS11, update disc min size requirements as of SAP Storage Guide.
Mandatory use npiv for scale-out.
2.3 10Gb Ethernet connectivity planning between AIX and Linux. HowTo instructions
(Dec/03/2015) added into the Planning Spreadsheet and the Implementation Guide
2.4 Higher core/memory ratio for BW workload on E870 and E880, FC-Port
(Jan/20/2016) requirements for scale out table update, min. SLES 11 SP4 Kernel=3.0.101-68-
ppc64
IBM Systems Solution for SAP HANA on POWER and IBM System Storage
3
Preface
SAP HANA on POWER is a new solution allowing customers to run HANA-based analytics and
business applications on a flexible IBM Power based infrastructure. Existing IT assets - servers,
storage, as well as skills and operation procedures - can easily be (re-)used, instead of enforcing
additional investment into dedicated SAP HANA only appliances.
About This Document

This document is intended for architects and specialists planning an SAP HANA on POWER
deployment. It describes the design considerations for hardware, networking, and software
components of the SAP HANA on POWER solution stack.
This guide does not replace existing SAP HANA documentation and sizing guides. It serves as a
supplement to the existing SAP HANA documentation and SAP Sizing methods to provide specific
guidance on how to meet SAP HANA sizing and operational requirements when running SAP HANA
on IBM Power Systems, IBM System Storage, IBM PowerVM, and Linux Operating System. It
describes the requirements for LAN and external SAN topologies.
IBM processes and contacts are introduced which help to obtain a valid configuration based on SAP
sizing for SAP HANA.
SAP Note 2055470 highlights all intermediate changes and provides an overview of the minimum
requirements as well as a reference to the most important SAP Notes.
IBM employees can access the ISICC SAP HANA on Power Systems community for up-to-date
materials complementary to this guide.
The most recent document version can be downloaded from IBM TechDocs:
http://www.ibm.com/support/techdocs/atsmastr.nsf/WebIndex/WP102502
Feel free to provide feedback and change requests for this document via email to isicc@de.ibm.com.
4
Table of Contents
Preface .................................................................................................................................................... 3
About This Document ............................................................................................................................. 3
Figures ................................................................................................................................................. 6
Tables .................................................................................................................................................. 6
Introduction ............................................................................................................................................ 7
SAP HANA on POWER Solution Overview ............................................................................................... 8
Hardware Planning................................................................................................................................ 10
SAP Sizing for SAP HANA on POWER................................................................................................. 11
Memory and CPU Capacity background information ................................................................... 12
Background information on core to memory ratio....................................................................... 12
Disk Quantification Background information ............................................................................... 13
Quick Reference: Get SAP HANA Sizing from SAP ........................................................................ 13
Links, References and Tools .......................................................................................................... 13
Planning Sheets: SAP Sizing........................................................................................................... 14
Mapping SAP Sizing Output to IBM Power Systems Configurations ................................................ 14
Planning Considerations for Power Systems Memory Capacity ................................................... 15
Planning Considerations for Power Systems CPU Capacity .......................................................... 15
Capacity on Demand ..................................................................................................................... 15
Virtualization / LPARs performance related considerations ......................................................... 15
Planning Considerations for Power Systems I/O Adapter Capacity.............................................. 16
Quick Reference: Find valid IBM Power Systems options ............................................................ 17
Planning Sheets: HANA Partition and IBM Power Server ............................................................. 19
Mapping SAP I/O KPIs to a SAN Storage Design ............................................................................... 20
Planning Considerations for a valid Storage Type ........................................................................ 20
Planning Considerations for SAP HANA Disk Volume and I/O Sizing ............................................ 21
Planning Considerations for SAP HANA scale-out installations .................................................... 22
Quick Reference: Find valid Storage Subsystem ........................................................................... 23
Planning Sheets: Storage .............................................................................................................. 24
SAP HANA Connectivity..................................................................................................................... 24
Planning Considerations for Virtual IPs......................................................................................... 25
Planning considerations for VIOS I/O virtualization ..................................................................... 25
Planning Considerations: Single-Host SAP HANA (scale-up) ......................................................... 25
5
Planning Considerations Multi-Host SAP HANA (scale-out) ......................................................... 26

Quick Reference: I/O Adapter Definition ...................................................................................... 27
Planning Sheets: Connectivity ....................................................................................................... 28
Software and Operating System ........................................................................................................... 29
Operating System.............................................................................................................................. 29
Planning Considerations ............................................................................................................... 29
Quick Reference: OS Planning....................................................................................................... 29
File System ........................................................................................................................................ 29
Planning Considerations for XFS with Multipathing ..................................................................... 30
Quick Reference: File System Definition ....................................................................................... 30
Planning Sheets: File System ......................................................................................................... 31
SAP HANA Software .......................................................................................................................... 31
Planning Considerations ............................................................................................................... 31
Quick Reference: SAP HANA Software.......................................................................................... 31
Links, References, and Tools ......................................................................................................... 32
Planning Sheets: Software and OS ................................................................................................ 32
Verification ............................................................................................................................................ 32
Optional Software and Hardware Considerations ................................................................................ 33
Migration .......................................................................................................................................... 33
Security ............................................................................................................................................. 33
Backup............................................................................................................................................... 34
IBM Spectrum Protect................................................................................................................... 34
SAN Infrastructure and SAP HANA File System Backup ................................................................ 34
SAP HANA Snapshots .................................................................................................................... 34
IBM Spectrum Scale and Elastic Storage Server ........................................................................... 35
High Availability and Disaster Recovery ............................................................................................ 35
SAP HANA System Replication (Multi-tier) ................................................................................... 35
SAP HANA Auto Restart, Auto Host failover, Failover Hooks ....................................................... 35
IBM Storage Replication................................................................................................................ 35
Management and Monitoring........................................................................................................... 36
PowerVC ........................................................................................................................................ 36
Support and Services ............................................................................................................................ 36
6
Getting help and information from the World Wide Web for IBM products ................................... 36
Getting help and information for IBM Server and Storage mapping ............................................... 36
Standard Support Flow ..................................................................................................................... 36
IBM Services, Support and Operation offerings for SAP HANA on Power ........................................ 37
Planning and Installation............................................................................................................... 38
Integrated Support........................................................................................................................ 38
Referenced documents ......................................................................................................................... 39
Copyrights and Trademarks .................................................................................................................. 41
Disclaimer and Special Notices ............................................................................................................. 42
Figures
Figure 1. SAP HANA on POWER Solution Stack....................................................................................... 8
Figure 2. Supported Application Server Environments........................................................................... 9
Figure 3. SAP HANA on IBM Power Servers Planning Process .............................................................. 10
Figure 4. Sample get_size report output .............................................................................................. 12
Figure 5. Quick Reference: "Get SAP HANA Sizing from SAP" .............................................................. 13
Figure 6. Quick Reference: Mapping SAP Sizing Output to IBM Power Systems.................................. 17
Figure 7. Floor and ceiling configuration for SAP HANA Business Warehouse partitions on IBM Power
Servers .................................................................................................................................................. 18
Figure 8. Quick Reference: Find valid Storage Subsystem .................................................................... 23
Figure 9. Minimum Single-Node SAP HANA Connectivity Setup with SAN Storage and Dual-VIOS ..... 27
Figure 10. Standard Support ownership of SAP HANA and Hardware Stack ........................................ 37
Figure 11. Support Flow for HANA on Power Solution Stack ................................................................ 37
Figure 12. IBM Services for SAP HANA on POWER environments ........................................................ 38
Tables
Table 1. SAP Sizing Summary ................................................................................................................ 14
Table 2. SAP HANA Partition Resource Planning Sheet ........................................................................ 19
Table 3. SAP HANA Server Planning Sheet ............................................................................................ 19
Table 4. Storage Planning Sheet ........................................................................................................... 24
Table 5. Minimum Single-Host SAP HANA I/O adapters for production .............................................. 26
Table 6. Minimum Multi-Host SAP HANA I/O Adapters per Server ..................................................... 26
Table 7. VIOS Partition Resource Planning Sheet ................................................................................. 28
Table 8. Ethernet Connectivity Planning Sheet..................................................................................... 28
Table 9. Fibre-Channel Planning Sheet ................................................................................................. 28
Table 10. SAP HANA File System ........................................................................................................... 31
Table 11. SAP Software Planning Sheet ................................................................................................ 32
Table 12. SAP HANA Installation Planning Sheet .................................................................................. 32
Table 13. Eco System Software for SAP HANA ..................................................................................... 33
7
Introduction
SAP HANA comprises several functional modules whose core is the SAP in-memory database. These
are deployed on top of a hardware and software stack, which should be planned according to this
Planning Guide.
Initially, SAP supported only the appliance delivery method for HANA in which certified hardware
partners deliver a HANA appliance. HANA TDI (Tailored Data Center) opened the appliance
deployment model to provide customers with more flexibility and choices. Initially only for the
storage infrastructure, but then also for the network and Server selection.
With the SAP HANA version for IBM Power Servers Architecture, SAP extended their offering once
more towards an enterprise hardware deployment model (TDI Phase 4). Customers can choose their
preferred IBM Power Server model - or virtualized partitions (LPARs, IFLs) of it - along with the best
matching storage components from a large menu of suitable hardware. By enabling customers to
reuse existing hardware and operation processes in their data centers, the SAP HANA on POWER
deployment model can significantly lower the costs and allow for easier integration of SAP HANA in a
customer data center.
A SAP HANA on POWER meets the requirements of customers for high flexibility in hardware
configurations and administration. SAP HANA systems can exploit PowerVM capabilities. This can
lower the entry barrier for SAP HANA significantly, since test systems can be easily installed on a
partition using available capacity or Capacity Upgrade on Demand resources. This eliminates the
need to purchase dedicated HANA hardware.
In case new server capacity to run SAP HANA has to be acquired, the IBM Power Systems Solution
Editions for SAP HANA offer attractive and flexible configurations (see PLET ZG15-0103).
From a technology perspective, IBM Power Systems are very well suited to provide excellent SAP
HANA database performance and reliability. SAP HANA on POWER exploits excellent memory
bandwidth, SIMD parallelization, and Simultaneous Multi-Threading (SMT).
Before starting, please download or check
SAP Note 2130682

SAP Note 2055470
SAP Note 218464 and 2227464
SAP Note 1943937
The SAP HANA on Power core documentation: http://help.sap.com/hana/
SAP HANA in Data Centers
Although this is not always aligned with the latest HANA SPS it gives a perfect link summary
and good introduction in pretested deployment options.
In cases where the SAP Note documents more current information than this Planning Guide,
the SAP OSS Notes supersedes this Planning Guide.
8
SAP HANA on POWER Solution Overview

Figure 1 shows the high-level solution stack1. On the infrastructure level existing hardware resources
can be used, assuming they provide sufficient capacity and fulfill the specified SAP performance KPIs
described later in this document.
Figure 1. SAP HANA on POWER Solution Stack
For SAP HANA in general, the required memory and compute capacity can be provided by either a
scale-out (multi-host) or scale-up (single-host) architecture. The supported servers and deployment
options (solution scope) are listed in the SAP central note for SAP HANA on POWER (SAP Note
2055470 and SAP Note 218464 or 2227464).
Note:
From the initial shipment stack (based on SPS9) to SPS10, SAP HANA on Power was a dedicated
product with limited HANA capabilities. Starting with SPS11 SAP ships a single SAP HANA product for
both Power and Intel.
The storage layer is designed to accommodate the different I/O characteristics of SAP HANA
persistent log and persistent data volumes. Another storage design aspect to be considered is to
allow for acceptable start-up times of a business critical HANA DB (RTO focused). Details can be
found in the chapter Mapping SAP I/O KPIs to a SAN Storage Design.
On operating system level SAP HANA on POWER is exclusively supported by specific Linux
distributions, as Figure 1 indicates. Details can be found in chapter Mapping SAP Sizing Output to
IBM Power Systems Configurations.
SAP product support and processes for the SAP HANA, SAP NetWeaver Business Warehouse (BW),
and SAP Business Suite layers will be consistent with the Intel-based platforms. The SAP Product
Availability Matrix (PAM) and related SAP Notes are the official sources to reflect this.
1
IBM offers system management tools focusing on the SAP landscape aspect what also includes the SAP
business application layer.
9
Figure 2 shows that IBM and many 3rd-party application server platforms can connect to a HANA on
POWER database instance.
Figure 2. Supported Application Server Environments
10
Hardware Planning
The following sections describe the hardware selection criteria and validation process based on an
SAP sizing along the workflow outlined in Figure 3:
Obtain required hardware capacity

requirements from SAP
Complete SAP HANA Apply SAP tools versus

Quicksizer process existing DB
SAP defined HANA Memory and Disk
configuration, SAPS
Identify Power System with sufficient DIMM and

adapter capacity
Calculate socket/core count
Enforced by # of DIMMs Match sizing output with SAPS capacity

table for cross check
Identify and Layout Storage System(s)
Meet RTO and RPO Match sizing output with

objective storage space
Map LAN & SAN I/O adapters into

selected system
OS System deployment
Verify SAP KPI-compliance of the

deployed system
HANA System deployment
Figure 3. SAP HANA on IBM Power Servers Planning Process
The orange part represents the initial sizing process as described in chapter SAP Sizing for SAP
HANA on POWER. It delivers some background information on the SAP HANA sizing methodologies
and tools. Further, it helps to identify the appropriate planning resources provided by SAP and IBM.
The blue parts refer to the IBM specific SAP HANA on POWER hardware design and layout steps
which are described in three chapters:
Mapping SAP Sizing Output to IBM Power Systems Configurations

Mapping SAP I/O KPIs to a SAN Storage Design
SAP HANA Connectivity
11
These chapters explain the background and provide the related IBM and SAP materials to select and
configure the hardware. It is important to consider these planning items as a whole in the context of
the existing SAP landscape.
The green parts represent the Operating System (OS) and SAP HANA software setup described in
chapter Software and Operating System.
Finally, the red box explains the verification of the deployed hardware and software stack. It is a
mandatory task to ensure production readiness and overall functionality of each individual customer
installation. This requirement is not specific to HANA on POWER deployments. Details and how to
ensure datacenter readiness using the SAP HWCCT (Hardware Configuration Check Tool) tool is
documented in the chapter Verification.
Support channels, IBM service offering details and contacts are summarized in chapter Support and
Services.
SAP Sizing for SAP HANA on POWER

IBM Power Systems are not sold as an SAP HANA appliance. Instead, requirements can be mapped
to any valid POWER processor-based system, as long as it provides the resources defined by the
following design considerations.
Sizing for SAP HANA on POWER is based on methodologies and tools provided by SAP SE. Their
output will specify vendor and platform independent SAP HANA system requirements.
System sizing for SAP HANA and its configuration is dominated by its physical memory demand. In
most cases the CPU capacity of a selected Power System is enforced by the memory size.
SAP HANA on POWER system configurations should not only focus on the SAP HANA database
instance itself, but also consider resources for SAP application servers and other SAP systems
maintained by the customer. Aggregating instance capacities and consolidating those with PowerVM
will establish a more efficient usage of IBM Power Systems. Always consider end-to-end solutions
throughout the design process also looking at other requirements like for example SAP HANA
Database resiliency and backup/restore.
Two basic SAP HANA sizing methods can be performed:
Inductive Method
Customer fills the SAP HANA relevant section of the IBM SAP Sizing Questionnaire for Suite and
BW with quantity structure of business KPIs. IBM TechLine transfers this input to SAP HANA
Quicksizer, which calculates system requirements in accordance with SAP rules.
In late 2014 SAP SE has released a new Quicksizer for HANA flavor to support this. The classical
Quicksizer (QS) version still is valid for sizing SAP Business Suite and other components.
In case SAPS are provided in the SAP HANA Quicksizer, they must be cross-checked with the ISICC
SAPS Capacity Tables for Power Systems to comply with the SAPS requirements specified during
the sizing step. For HANA make sure to use the Linux on POWER sizing tables.
12
Today, the HANA Quicksizer specifies SAPS values for Suite on HANA scenarios; however, it does not
for NetWeaver Business Warehouse. The results are the required server memory and disk space. In
order to calculate processor counts please contact the IBM TechLine and read SAP Note 2055470.
Deductive Method
Existing SAP Business Warehouse users intending to migrate to SAP HANA can run a SAP provided
report within their existing BW system. SAP Note 1736976 Sizing-Program for BW on HANA
describes how to use the get_size report. SAP SE strongly recommends using this method as it
provides more accurate and detailed sizing input.
Source SAP BW Table Sizes
Target SAP HANA DB Size
For scale-up these values

have to be aggregated onto a
single node
More Details (growth etc.)

Figure 4. Sample get_size report output
Memory and CPU Capacity background information

SAP publishes a good document Sizing Approaches for SAP HANA Lessons Learned. On the one
hand it gives a comprehensive background and on the other it provides basic rules to conveniently
come up with an initial capacity assessment for customer opportunities.
The ISICC SAPS Capacity Tables for Power Systems contain a Maximum Memory column to
identify maximum possible system memory in GB, dependent on number of sockets.
Background information on core to memory ratio

For BW the SAP HANA core to Memory ratio defines the maximum amount of Memory a single core
can drive under highest load conditions under BW-EML workload. Less memory per core is allowed.
For exceptions please contact your SAP and IBM representative to get a list of possible options.
The core to memory ratio as well as the verified floor and ceiling CPU and Memory requirements are
defined in SAP Note 2188482.The existing values cannot be extrapolated to larger systems. The
ranges have to be put into the context of the customer specific workload as they are based on a BW-
EML benchmark.
13
Disk Quantification Background information

The SAP HANA Server Installation Guide, the SAP HANA Storage Requirements Guide, along
with the result of the SAP HANA Quicksizer gives a good understanding of the required volume
requirements of storage. The sizing output itself is not sufficient.
Quick Reference: Get SAP HANA Sizing from SAP

Make yourself familiar with HANA Sizing basics to be able to complete the
SAP Sizing Questionnaire for Techline or the SAP HANA Quicksizer Tool
(see: Links, References and Tools subchapter)
Need yes contact your respective TechLine, FTSS

support ? and/or ATS team
Business yes Apply SAP tools versus existing DB

Warehouse (SAP-Note 1736976)
Complete SAP Quicksizer process or use the SAP Sizing

Questionnaire from Techline (www.ibm.com/erp/sizing)
Quantified system requirements (memory GB, cores, SAPS, volume)

available from SAP.
Figure 5. Quick Reference: "Get SAP HANA Sizing from SAP"
Links, References and Tools

Sizing Approaches for SAP HANA Lessons Learned
How to size SAP BW on SAP HANA
Sizing for SAP Business Suite powered by SAP HANA
IT Planning Documents (SAP Wiki)
SAP Sizing Questionnaire for Suite and BW
Quick Sizer Tool
Read more about SAP HANA sizing on SAP Community Network (Registered S-Users only)
Central SAP Notes about SAP HANA sizing:
Central BW Sizing SAP Notes: SAP Note 1637145 and SAP Note 1736976
Central SAP Suite Sizing SAP Notes: SAP Note 1793345 and SAP Note 1872170
Central SAP HANA Sidecar Solution SAP Note: SAP Note 1514966
SAP Note 1599888 - SAP HANA: Operational Concept
14
Planning Sheets: SAP Sizing

Table 1. SAP Sizing Summary
No SAP Planning Sheet SAP Sizing Result Notes

1 Memory for the SAP HANA DB See SAP HANA Quicksizer output
2 SAPs/Cores for SAP HANA See SAP HANA Quicksizer and
DATABASE Planning Guide
3 SAP HANA disk space for data See SAP HANA Quicksizer output
4 SAP HANA disk space for log See SAP HANA Quicksizer output
5 SAP HANA disk space for share See SAP HANA Quicksizer output
6 Minimum Storage KPIs See SAP Note 19439372
7 Number of HANA nodes e.g. get_size report3
Table 1 is the basis for the other tables. Values inserted here build the absolute minimum
requirement for this SAP HANA instance and will in most cases not be sufficient for the final solution.
Note that additional requirements for Virtual I/O Servers (VIOS), application servers or other
applications and multiple-LPAR deployments in one server will increase the overall requirements of
resources.
Mapping SAP Sizing Output to IBM Power Systems Configurations

The server decision has to be made to fit the memory, CPU and I/O adapter requirements including
the planning of additional workload running on the same server.
For scale-up3 (single SAP HANA node) all must fit into a single POWER processor-based server.
For scale-out (multi-host) the inter-LPAR ethernet communication, the auto-host-failover

architecture (shared disc vs. shared filesystem) must be considered when planning the server as this
includes additional connectivity aspects. In IBM Power scale-out setups can comprise out of multiple
LPARs in a single server as well as spanning multiple servers. In both case the rule of thumb is to first
scale-up and then scale-out.
Sizing for SAP HANA System Replication allows the standby server to run additional workload when
setup in the so called cost optimized setup. For more details read the SAP HANA documentation or
contact your cluster vendor automating the SAP HANA SR if predefined solutions exist (link to the
IBM cluster solution).
To optimize TCO according to SLA criteria, you must differentiate between mission critical
production and less critical non-production SAP HANA database servers (see also
https://blogs.saphana.com/2014/08/12/cost-optimized-sap-hana-infrastructure-requirements-for-
non-production-usage/).
2
SAP specifies latency and MB/s. To transfer MB/s to IOPS: MB/s * 1024 = IOPS
3
For scale-up setups the number of SAP HANA nodes can be multiplied with the resources to plan for a single
partition setup.
15
The chosen Power Server has to fulfill all requirements (including e.g. adapter slots for I/O. Refer to
the Power Systems Facts and Features on Sales Support Information (SSI) or similar documents for
detailed IBM Power Systems specifications(for example POWER8 Facts and Features).
Linux only server models are available and can be an alternative at a compelling price. The same
applies to Integrated Facility for Linux (IFL) specialty processors which are dedicated to run Linux
operating systems only in the E-Class models. For the E-class models these allow tailoring IFL
memory to host a SAP HANA DB.
An effective way to accomplish a HANA configuration is based on the special priced IBM Power
Systems Solution Editions for SAP HANA.
Note:
SAP HANA server planning comprises of three parts:
The size of the SAP HANA partition

= SAP HANA Sizing + SLA requirements
Adapter/storage planning as outlined in the next chapters
= SAP TDI/Ethernet sizing + VIOS + SLA requirements
The size of the IBM Power Server running multiple partitions next to HANA
= workload consolidation + VIOS
Planning Considerations for Power Systems Memory Capacity

PowerVM AME (Advanced Memory Expansion) and AMS (Advanced Memory Sharing) is NOT
supported for SAP HANA database partitions. This is independent of future support by Linux
distributions.
Planning Considerations for Power Systems CPU Capacity

SAP Note 2188482 documents the core/memory ratio for IBM Power Servers for Business
Warehouse and Business Suite. This ratio is linked to a workload pattern which assumes the queries
constantly require all memory to be accessed (worst case scenario). It has to be ensured that the
context of the ratio (workload and core to memory ratio) is mapped to the actual deployment and
business need. If the planned system size does not fit in the given ratios please read chapter
Background information on core to memory ratio.
Capacity on Demand
In the SAP HANA, version for IBM Power Servers Architecture Initial Shipment Stack up to SPS11
CoD planning is possible for a subset of possible architectures. The SAP HANA code requires a cold
operation to enable HANA to pick up the added or removed resources (restart). Ensure that the ini-
parameter files are configured appropriately. This can be achieved by installing HANA once with full
capacity and once with reduced capacity and compare the *.ini files of HANA. Before production
usage a SAP review has to be requested to approve the solution.
Ensure to verify your Linux operating system already supports CoD operations transparently. The
first SLES to support core CoD functionality is SLES 11 SP4.
Virtualization / LPARs performance related considerations

In order to minimize Hypervisor dispatching latency and to provide best memory affinity, only
dedicated or dedicated donating LPARs are supported for SAP HANA productions partitions. The
16
dedicated donating partition feature grants the performance benefit of dedicated resources without
the waste. It allows for the donation of spare CPU cycles of dedicated processor partitions to be
utilized by the shared pool, thus increasing the overall system performance. The dedicated partition
maintains absolute priority for dedicated CPU cycles, and sharing will only occur when the dedicated
partition has not consumed its resources.
Relaxed conditions for all none-production LPARs apply as of SAP Note 2055470 - HANA on POWER
planning and installation specifics - central note.
SAP HANA is highly optimized for the underlying hardware in order to achieve its superior
performance. Those optimizations include sophisticated algorithms using processor features (like
vector facilities) as well as using deep knowledge on the underlying cache structures. Hence, the
partitions on IBM Power Systems should be created in a specific manner to allow best performance.
Following you will find a set of recommendations to achieve a good partition layout for SAP HANA:
SAP HANA benefits especially on larger systems from an even distribution of compute power
across the LPAR. PowerVM will automatically achieve this, when the number of CPUs is the
same for each socket.
The LPAR layout can be verified with the command numactl --hardware and should show the
same number of CPUs for each numa node.
When dynamic operations are planned (such as CoD or LPM) the numa layout has to be
ensured also at the target if the HANA instance needs to provide the same performance as
before. (Valid for SAP HANA Version SPS9 SPS11)
Especially on larger systems with multiple partitions, which have been created and deleted
over time, PowerVM might not be able to achieve the best partition layout immediately. In
this case you can use DPO (Dynamic Platform Optimizer) to change the LPAR layout. This can
be done on-line without down-time of the SAP HANA system.
For customers planning for a HANA workload with high performance requirements always
choose the high frequency server model.
Planning Considerations for Power Systems I/O Adapter Capacity

Please see chapter SAP HANA Connectivity to ensure the server provides sufficient adapter slots.
When using VIOS ensure an appropriate resource planning for these LPARs as a minimum of 10Gb
Ethernet cards is required.
17
Quick Reference: Find valid IBM Power Systems options
is this server for

production HANA
LPARs? no
yes
For none-production partitions
Quantified system requirements relaxed Storage, Server, core and
available from SAP HANA Sizing memory requirements apply.
The SAP Software must be used
ONLY for the released functional
scope part of the license.
Indentify IBM Power Server options providing enough memory as sized by SAP
(see POWER Systems Facts & Features or SSI for Server details.
Verify that only SAP HANA approved Power Hardware as of SAP Note 2133369 is used)
yes
SAPS
available
Verify SAPS requirements can be met
(see: ISICC SAPS Capacity Tables for no
Power Systems for LINUX)
yes core/memory
ratio available
Verify the partition is sized to not exceed the
amount of memory a single core can drive (SAP No (all non-prod LPARs)
Note 2133369 and 1903576 for extensions)
Add additional requirements

Ensure later the Server can also
for additional workload
fulfill all I/O Adapter Requirements
(e.g. VIOS, Appserver, ...)
List of valid IBM Power Servers available.
Figure 6. Quick Reference: Mapping SAP Sizing Output to IBM Power Systems
In Figure 7 the LPAR configuration options for a Business Warehouse SAP HANA are outlined. For
production HANA partitions there is a range of available partition sizes in any IBM Power Server with
POWER8 processor architecture (middle and right hand partition/server options). The partition must
be between the floor and ceiling and fulfill the core to memory ratio. In scale-out multiple of these
LPARs are used.
To reflect the POWER architecture and different workload patterns a SAP HANA can run, SAP has
established a process to extend the options provided (yellow box with red letters on the right top).
18
Any IBM Power Server with Any IBM Power Server with Any IBM Power Server with
min. POWER7+ min. POWER8 min. POWER8
Processor Architecture Processor Architecture Processor Architecture
Other core/memory ratios:
SAP Notes
1903576 + Contact your
SAP/IBM representative for
Other Other additional options
Other
workload
optional: Other
workload
optional: HANA Partition
workload
other workload
other
workload workload Max
4,8 TB Memory
Other
and workload
Other
optional:
workload
Max 96 Cores other
HANA workload
Partition per LPAR
Select partition
HANA Partition Min. 8 size within this
Cores range. Memory
Recommended as per core to
64 GB Memory
memory ratio
2 Cores
Sandbox and non-production Floor configuration eligable to Ceiling configuration eligable to

HANA installations run production systems. run production systems.
Check SAP Note 2055470 for (Oct/2015) (Oct/2015)
updates. Check SAP Note 2188482 for
updates.
Figure 7. Floor and ceiling configuration for SAP HANA Business Warehouse partitions on IBM Power
Servers
Please verify SAP Note SAP Note 2188482 - SAP HANA on IBM Power Systems: Allowed
Hardware for detailed Production requirements.
Please verify SAP Note 2055470 for relaxations for non-production systems
Note:
If there are more than one viable Server always prefer the one with the highest frequency. For
example the E850 is allowed to be used up to 32 cores what is eligible to run the high frequency
model for performance critical installations.

ISICC SAPS Capacity Tables for Power Systems (IBM only):
Use the table to identify systems and number of cores. The SAPS tables are maintained as an IBM
Notes database and only accessible for IBM employees. BPs should contact IBM TechLine for
support. This is identical to sizing systems for other SAP workloads.
Refer to the IBM Power Systems Facts and Features for detailed system specifications (IBM only):
POWER8 Facts and Features

POWER7+ Facts and Features
Sales Support Information (IBM only)
19
Planning Sheets: HANA Partition and IBM Power Server

Table 2. SAP HANA Partition Resource Planning Sheet
No Partition Planning Sheet # per HANA partition Sum Notes
Number of dedicated/~
1 Table 1.7
donating SAP HANA LPARs
2 Memory per Partition Table 1.1
IBM TechLine mapping based on

3 CPU/SMT per Partition
SAP sizing (Processor specific)
# Ethernet paths seen in

4 Defines minimum for Table 8
Partition
# active FC paths seen in

5 Defines minimum for Table 9
Partition
6 Partition IPs/hostnames
Table 3. SAP HANA Server Planning Sheet
No Server Planning Sheet Value Notes

1 Power Server Model(s) See SAP Note 2055470
2 Number of Servers for high availability min 2 Servers
3 SAP HANA Memory ++ Tables:
Overall Server Memory 2.2+7.6+3.8+13+hypervisor +/-(spare|CoD)
resources
4 SAP HANA CPU ++ Tables:
Overall Server CPU 2.2+7.5+3.8+13+hypervisor +/- (spare|CoD)
ressources
5 Consider SR-IOV capable ports for VIOS
Number of Ethernet I/O
deployments, depends on VIOS planning, Tables:
Adapter/Ports
8.6+8.7+7.3+3.8+13+spare-capability of sharing
6 Consider NPIV capable Ports, depends on VIOS
Number of FC I/O planning, Tables: 7.4+9.1+9.5+3.8+13+spare-
Adapter/Ports apability of sharing, verify if the number of paths
seen inside the partition can be achieved.
7 VIOS planning desired (yes/no) See Table 7. VIOS Server Resource Planning Sheet
8 Additional workloads planned? Provide location of documentation
9 PowerVM, see SAP note 2055470 for supported
Virtualization Technology used
alternatives
If a new server is deployed, keep in mind that processors and memory are required for Virtual I/O
server partitions (VIOS). Power Hypervisor (PHYP) also uses memory at a max of 8% of total server
memory. Do not under estimate VIOS sizing as it can hurt SAP HANA performance.
20
Mapping SAP I/O KPIs to a SAN Storage Design

SAP HANA database performance depends on processing everything in memory. Hence the ongoing
load and store operation are reduced to initial loads at first access of data and writes to log files and
save-points. The requirements introduced by the SAP HANA I/O characteristics for the file system
and the underlying storage have four planning objectives:
Start-up time:
Starting an SAP HANA Database from the stored content on disk requires read performance
during startup.
Database Persistency:
To provide persistency, data and log content is written regularly to disk. This requires low
latency especially for log volumes. SAP also offers to hold multiple versions of the data area,
multiply the storage requirements if used.
Backup:
SAP HANA snapshots freezes the data file system for a period of time. This can be seconds,
when using FlashCopy storage snapshots, or up to minutes doing a file system copy in the
OS. The faster the write to the backup, the sooner the SAP HANA database can resume
writing save-points again. SAP HANA database operations will continue. Find more
information about backups in chapter Backup.
Data Protection:
A storage subsystem can provide functionality to duplicate data to a second location. This
can be a storage mirror; IBM Spectrum Virtualize (SVC) stretched cluster mdisc mirroring,
Hyperswap or similar storage methods to duplicate data to a second location.
To optimize TCO according to SLA criteria, you must differentiate between mission critical
production and less critical non-production SAP HANA database servers (see also
https://blogs.saphana.com/2014/08/12/cost-optimized-sap-hana-infrastructure-requirements-for-
non-production-usage/).
Planning Considerations for a valid Storage Type

Although every storage subsystem fulfilling the SAP KPI requirements is eligible to run SAP HANA on
POWER, we recommend focusing on TDI certified systems.
1. Storage systems having passed the test are published on the SAP Certified Enterprise
Storage Hardware for SAP HANA web-site.
2. Ensure appropriate multipath drivers do exist in the targeted Linux Operating system and
used Power technology (e.g. NPIV).
3. The end-to-end setup up to the server comprising of filesystem, operating system, server,
network and storage needs to be tested with the SAP HANA Hardware Configuration Check
Tool HWCCT.
4. Besides fulfilling the SAP minimum I/O requirements the setup must allow to load the full
data into memory accordingly to the customer RTO requirements for initial startup and SAP
HANA Host Auto-Failover feature (available for scale-up and scale-out).
5. For large SAP HANA scale-out deployments verify if the maximum allowed number of
multipath paths per OS is sufficient when deploy the data and log discs based on a shred disc
approach. With SVC technology (Storewize, ) each device seen in the multipath driver
21
shows n active and n passive paths. Not using SVC technology removes the paths used for
the passive so called enabled paths what divides the number of paths by 2.
6. Storage subsystems can be attached either natively or via a storage virtualization layer like
IBM Spectrum Virtualize (formerly known as IBM SAN Volume Controller, SVC).
7. Redundant SAN attachments via virtual or dedicated fibre-channel (FC) adapters are possible
(VIOS).
Planning Considerations for SAP HANA Disk Volume and I/O Sizing
Use following SAP documents as a starting point: SAP HANA Storage Requirements and SAP
HANA Server Installation Guide.
Note: Be aware that the mapping of the SAP sizing output and the following planning steps identify a
larger disk space requirement than the sizing provided by the SAP HANA Quicksizer and the SAP
HANA Storage Requirements paper, based on the desired backup, advanced performance and
resiliency capabilities. However, the baseline requirements are identical to the TDI Storage
Requirements given by SAP.
The SAP HANA storage capacity requirements consist of several elements with different
characteristics:
SAP HANA install directory and /usr/sap

local directory
SAP HANA share
A shared filesystem spanning all SAP HANA nodes belonging to the same system. It holds the
SAP HANA executables.
SAP HANA data
Performance relevant for fast RTO times on startup (mainly read). Needs to fulfill the
minimum Storage KPIs validated by HWCCT. Runs a supported Filesystem for SAP HANA as
documented by SAP.
SAP HANA log
Benefits from low latency I/O characteristics (mainly write). Needs to fulfill the minimum
Storage KPIs validated by HWCCT. Runs a supported Filesystem for SAP HANA as
documented by SAP.
Number of online SAP HANA backups
Typically 0-3 with same characteristics as SAP HANA data
The Operating System
IBM developed the SAPmagic - Storage Sizing Tool for SAP landscapes, which can be used to verify
or outline options complementing the SAP HANA Quicksizer across IBM storage technologies for the
SAP HANA data, log and shared filesystems. It includes a special method for optimizing SAP HANA
file placement and I/O bandwidth leveraging a balanced blend of HDDs and SSDs/Flash.
See the IBM System Storage Architecture and Configuration Guide for SAP HANA TDI on
TechDocs for a detailed description of SAPmagic methods.
22
Note:
SAPmagic comes with no guarantee. Some aspects for a POWER based deployment are not covered
as it is created for the appliance server TDI deployments. But it is a good start point and can be well
used for verification.
Additional Sizing Considerations for Backup, DR and HA

An SAP HANA database landscape also requires space and functionality for:
SAP HANA file system backups and snapshots.

Disaster recovery data protection.
High availability data protection.
SAP HANA backups consist of an individual strategy for SAP HANA data, log, and
configuration files. The SAP HANA Administration Guide provides all required information
to determine the space and setup requirements and the different options for creating the
backup. Some of the SAP Solutions do rely on storage subsystem features such as FlashCopy.
To protect the data in case of a disaster, the SAP HANA database content can be replicated on
database level using SAP HANA System Replication or a storage subsystem mirror to a secondary
site. The mirroring technology must guarantee a logical order in sense of I/O requests. Details can be
found in the Introduction to High Availability for SAP HANA document from SAP.
Planning Considerations for SAP HANA scale-out installations

SAP enforces to install scale-out HANA instances with Auto-Host-Failover. Even for SAP HANA System
Replication enabled landscapes. Verify with SAP starting with SPS12 if this is changed.
Auto Host Failover requires enabling the hosts to access the data volumes of a different host in case
of a failover. This can be done either by a shared filesystem (verify SAP Note 20555470 for supported
options) or by shared disks and scsi-reservations. Please read the SAP HANA Server Installation
Guide and the referenced documents for LUN granularity and required host attachments. As an
intermediate Solution until the Server Installation guide ships all required steps IBM provides a quick
guide on IBM Techdocs.
For XFS the shared disc deployment is mandatory. Except for Lenovo this is the most common
deployment style for SAP HANA scale-out deployments. This requires other preparation tasks
compared to shared filesystems. The following is a list of important things for the planning:
All worker nodes get a dedicated XFS filesystem for data and log
/etc/sudoers must remain to enable the fcClient(LVM) to work
The LUNs for data and log have to be setup as a shared disc on the SAN, attached to all
worker and standby nodes.
fcClientLVM is the SAP HANA build in tool for disc reservation handling and auto-mounting
the filesystems on the failover node. This requires before the installation to prepare the
global.ini file as documented by SAP.
fcClientLVM only works for npiv, not for vscsi. Ensure to setup the VIOS Server appropriately.
23
Quick Reference: Find valid Storage Subsystem

Customer SLAs:
SAP minimum requirements:
Max startup time of HANA
SAP HANA Storage KPIs for TDI
Backup, HA and DR requirements
SAP volume requirement based on the SAP
Requirements from other workload
HANA sizing effort
(Appserver, VIOS, ....)
Identify required Storage features (OS multipath

drivers, HA, DR, redundency and backup)
Calculate overall disk capacity requirement for HANA

by SLA quality (SSD/Flash vs. Disk)
For IBM storage subsystems :

Identify valid storage options based on capacity and
I/O performance using SAPmagic tool for HANA
Ensure later the I/O Adapter Add additional requirements (e.g.

Requirements are met as well. VIOS, SAP Appserver, ...)
List of valid storage subsystems for the SAP HANA

workload available.
Figure 8. Quick Reference: Find valid Storage Subsystem

SAP Documentation:
SAP HANA Storage Requirements
SAP Certified Enterprise Storage Hardware for SAP HANA
SAP HANA Administration Guide
Introduction to High Availability for SAP HANA
SAP Note 1943937 - Hardware Configuration Check Tool - Central Note (defines also the SAP KPIs)
SAP Active Global Support offers the SAP HANA Going-Live Check which among other tests
conducts a data throughput test using the SAP HANA Hardware Configuration Check Tool.
IBM Documentation:
IBM System Storage Architecture and Configuration Guide for SAP HANA TDI
Tools:
Quick Sizer Tool
SAPmagic - Storage Sizing Tool for SAP landscapes
24
Planning Sheets: Storage

Table 4. Storage Planning Sheet
No Storage planning Sheet Description Notes

1 IBM Spectrum Virtualize
E.g.no, yes or Stretched
configuration (SVC)
2 #/Storage Type/Model/Purpose e.g. 1/Flash System/V840/SAP HANA Log
3 #/Storage Type/Model/Purpose e.g. 2/XiV/Gen4/all except SAP HANA Log
4 IP(s) for Storage/Login Consider redundant access for HA purposes,
Credentials document where to obtain credentials
5 e.g. Storage Mirroring, FlashCopy, . See
Required Storage licenses also Optional Software and Hardware
Considerations
6 Additional Space for Backup,
HA,
7 Overall Space requirements / Calculated from Tables: 10 (# of HANA
Storage System partitions)+7.7+3.8+4.6+13+spare resources
8 # of cables for throughput
>3
requirements
SAP HANA Connectivity

The I/O requirements are identical to those of Intel based SAP HANA systems with external storage.
However, for SAP HANA on POWER landscapes, these adapters can either be installed as dedicated
or shared (VIOS) adapters. With PowerVM, additional workload can be put on the same server which
introduces additional requirements for the overall planning.
The design of the SAP HANA node connectivity needs to reflect zoning, security, and throughput
considerations. In particular aggregated fiber channel bandwidth needs to cope with the throughput
requirements of the storage layout designed in the previous step.
All I/O adapters have to be configured redundantly in order to meet resiliency requirements of a
production enterprise database. For shared adapters this always implies a dual-VIO server setup.
Adapter requirements for an SAP HANA instance look different for single-host (scale-up) and multi-
host (scale-out) SAP HANA systems.
SAP HANA network planning need to consider bandwidth requirements for the following
communication paths4:
Communication - LAN
o SAP HANA database tier to SAP application server tier
o SAP HANA inter-node communication (scale-out only)
o SAP HANA System Replication Communication
o SAP HANA share (for CTS+ and/or scale-out)
o Administrator network
Storage SAN
4
This does not take GPFS planning into account.
25
Planning Considerations for Virtual IPs

Similar to SAP NetWeaver systems, a SAP HANA database can be installed using a virtual IP address.
Beside the standard of using virtual IPs for SAP applications there are two cases where a virtual IP for
the SAP HANA Database becomes mandatory:
SAP management tools such as SAP Landscape Virtualization Management (LVM) rely on
virtual IPs to operate on the SAP instances.
Most cluster solutions require having a virtual IP to failover a SAP HANA System Replication.
SAP HANA itself also provides such capabilities.
SAP Note 962955 and SAP Note 1900823 provide additional information
Planning considerations for VIOS I/O virtualization

VIOS is a strategic component for IBM and most customers deploying partitions on IBM Power
Systems. VIOS based deployments are mandatory for consolidation, Live Partition Mobility (LPM5)
and other features.
For fibre-channel virtualization, NPIV capable adapters should to be chosen to take advantage of
functionality in the larger eco-system and ease of use.
For Fibre Channel in multipath environments consider:
Each LUN should be accessible through 4 paths to utilize striping in XFS.
For Ethernet consider:
I/O sensitive internode communication does not exist for scale-up. Using the full 10Gb
Ethernet bandwidth is not critical for scale-up deployments.
SR-IOV capable adapters are to be evaluated for Ethernet connectivity. For traditional
Ethernet virtualization plan for appropriate CPU sizing and dedicated donating cores.
Jumbo frames with a MTU size of 9000 are required for native and VIOS attached Ethernet
adapters to get reasonable throughput. This has to be enabled end-to-end including the
network switches. Not enabling this will limit the Ethernet I/O to 2-3Gb/s.
VIOS in a mixed environment using AIX and Linux require having same large_send and
large_receive settings end-to-end when using the same infrastructure/VIOS adapter. For
using large_receive in a Linux environment Linux must support large_receive (starting with
SLES11 SP4) and the VIOS version must contain iFix IV72825 or the VIOS version number
must be >= 2.2.4.0.
SUSE SLES 11 SP 4 kernel 3.0.101-68-ppc64 has the large_send/large_receive enablement
what brings advanced throughput. Verify support when used across different Operating
Systems.
Planning Considerations: Single-Host SAP HANA (scale-up)

As part of the Power System configuration process, adapters need to be selected that contain
sufficient ports of the specified connection speed. These need to fit into available PCI slots of the
5
Considering the large memory footprint LPM can only be done under load conditions allowing the memory to
be transferred in a reasonable amount of time.
26
planned server.
Additional adapters should be considered when planning for a later multi-host landscape for growth
or resiliency reasons or additional workload (see subsequent chapter).
Minimum I/O adapter requirements for a production SAP HANA system using dual-VIOS are6:
Table 5. Minimum Single-Host SAP HANA I/O adapters for production (floor configuration as required by SAP)
VIOS #1 VIOS #2 Function

LAN / 1* 10 Gb port 1* 10 Gb port SAP HANA database to
Ethernet app server
1* 1 Gb port 1* 1 Gb port Administrator and SAP
HANA share network
Peak loads up to 300MB/s Peak loads up to 300MB/s SAP HANA System
per HANA node.7 per HANA node.7 Replication (optional)
SAN / fibre- 2 * 8Gb ports 2 * 8Gb ports Attachment to data and
channel log persistency
The network ports shown in Table 5 are a starting point and need adaptation for the specific
installation. The pattern can be mapped into any supported Power System which fulfills previous
sizing criteria and has sufficient free adapter slots. It makes no difference whether this is applied to a
dedicated SAP HANA on POWER server or to a partition in a larger consolidation system.
Planning Considerations Multi-Host SAP HANA (scale-out)

Note: SLES 11 SP4 has to be installed as described in the IBM Implementation guide.
The members in a scale-out setup need to coordinate their workload during a query. This requires
additional dedicated LAN segments as per SAP SE design guideline.
Minimum I/O adapter requirements for a production SAP HANA system using dual-VIOS are:
Table 6. Minimum Multi-Host SAP HANA I/O Adapters per Server (floor configuration as required by SAP)
VIOS #1 VIOS #2 Function

LAN / 1* 10 Gb port 1* 10 Gb port SAP HANA database to
Ethernet application server
1* 1 Gb port 1* 1 Gb port Administrator network
2* 10 Gb port 2* 10 Gb port SAP HANA inter-node
connect
<under investigation> <under investigation> GPFS8 network
(optional)9
6
When using dedicated adapters, each HANA node requires the same amount of adapters as a dual VIOS setup
would require. However, if a HANA node uses the entire server, native attachment is a reasonable option if no
VIOS features (e.g. Live Partition Migration (LPM)) are used.
7
The peak load documented is the highest known number. Campus solutions where both sides are
synchronously replicated to avoid data loss in a HA fail-over scenario have to deal with the individual peak
loads. DR setups between sides typically have an intentional delay between tiers and by that the need to cover
peak loads is lower.
27
Peak loads up to 300MB/s Peak loads up to 300MB/s SAP HANA System

per HANA node.7 per HANA node.7 Replication (optional)
SAN / fibre- 2 * 8Gb ports10 2 * 8Gb ports Attachment to data and
channel log persistency (SAN)
(HWCCT validated for 4
production HANA LPARs.)
Same as in previous paragraph applies for mapping these ports to proposed hardware.
Frontend Single Node SAP HANA Server
Dedicated[-donating] Partition for HANA
ETH0 ETH1 FC0 FC1 FC2 FC3

SAP, IBM, 3rd party
management and monitoring
tools
Dedicated Dedicated
Partition Partition
VIOS VIOS
SAP application server

1 10 HBA HBA 10 1
GB GB GB GB
Admin
FC Switch FC Switch
SAN Storage
Business
Business
Application
Application
Figure 9. Minimum Single-Node SAP HANA Connectivity Setup with SAN Storage and Dual-VIOS
Quick Reference: I/O Adapter Definition

Calculate the fibre-channel throughput to meet the desired start time to load the data into memory.
This value must not be smaller than the SAP provided KPIs. The client and SAP HANA node
communication requirements are defined by SAP in the TDI deployment guides.
8
Support for GPFS is handled by IBM PMRs. GPFS in SAP HANA on POWER deployments is not pre-evaluated
for SAP HANA data and log files and by that no direct SAP solution support is granted.
9
The requirements may differ compared to Intel deployments. Details will be provided along with the GA of
the scale-out solution.
10
The minimum sizing for scale-out and scale-up is identical. For scale-up it is driven by the need for
redundancy, for scale-out or multiple LPAR support it provides sufficient I/O capability based on the
measurements performed for the multiple LPAR support (4 production systems) described in SAP Note
2230704.
28

IBM System Storage Architecture and Configuration Guide for SAP HANA TDI
SAP HANA Network Requirements
Planning Sheets: Connectivity

The planning sheets below can be used to plan and document the SAP HANA environment in a
Power infrastructure. The planning has to be done for each Server unless all are identical.
Table 7. VIOS Partition Resource Planning Sheet
No VIOS Partition planning sheet I/O virtualization Notes

1 Min. Version 2.2, dual VIOS, SEA
Failover or newer technology. If
VIOS Release 10 Gb Ethernet is used min.
Version 2.2.4.0 or 2.2.3.x with iFix
IV72825.
2 Dedicated partitions are
VIOS Partition Type
recommended. (2* for dual setup)
3 Ethernet Virtualization See Table 8
4 FC Virtualization NPIV for MPIO See table 9
5 Min 1, for large production min. 2,
CPU for each VIOS
none-prod shared pool is possible
6 Memory for each VIOS Min 4GB
7 Disk space for each VIOS
Table 8. Ethernet Connectivity Planning Sheet
No Ethernet planning Sheet Zone Adapter/Speed/# Notes

1 Only for
SAP HANA network SAP HANA network
scale-out
2 Admin network Admin network
3 Appserver connect Appserver connect
4 Typically
SAP HANA share NFS or SAP HANA share
GPFS
5 GPFS (Eth or IB) GPFS (Eth or IB)
6 Sum
Sum HBA/Port/Zone
HBA/Port/Zone
HANA
HANA
7 Sum
Sum HBA/Port/Zone
HBA/Port/Zone
Table 3.8+13
Table 3.8+13
Table 9. Fibre-Channel Planning Sheet
No FC Planning sheet per Server Sizing Notes

1 Ports/HBA for HANA Min 4 Ports on 2 HBAs
2 Switch Model/Speed
3 Switch/Zone 1
4 Switch/Zone 2
5 Ports/HBA for Table 3.8 and 13
29
Software and Operating System

The Linux setup of SAP HANA on POWER does not differ in essence from any Intel SAP HANA
deployment. The central SAP document to follow is the SAP HANA Server Installation Guide.
Operating System
The basis is the same as for all other SAP supported Linux distributions. Ensure appropriate licensing,
versions, distributions and configurations as documented in SAP Notes listed in the Links,
References and Tools section in this chapter.
Planning Considerations
Select any documented Linux distribution (SAP Note 2055470). Do not deploy SLES 11 SP3 anymore.
If SLES 11 SP3 is already installed it is strongly recommend to plan for an update to SP4 with a
minimum kernel of 3.0.101-68-ppc64.
Download the SAP HANA SUSE installation guide (SAP Note 1944799 - SAP HANA Guidelines for SLES
Operating System Installation).
Use the SAP HANA on IBM Power Systems and IBM System Storage (Supplemental IBM
Installation Guide to the SAP HANA Master, Implementation and Administration Guides) in
addition to above guides for IBM specifics.
Quick Reference: OS Planning

Refer to the SAP manuals to setup Linux for SAP HANA and the SAP HANA on POWER
Implementation Guide. Both are referenced in the subchapter Links, References and Tools. In
addition appropriate OS license and install media have to be in place.

Installation Guides:
SAP HANA Server Installation Guide
SUSE LINUX Enterprise Server Deployment Guide
SUSE Knowledge Base
SAP HANA on IBM Power Systems and IBM System Storage (Supplemental IBM Installation
Guide to the SAP HANA Master, Implementation and Administration Guides)
SAP Notes:
SAP Note 2055470 - HANA on POWER planning and installation specifics - central note
SAP Note 1944799 - SAP HANA Guidelines for SLES Operating System Installation
SAP Note 1310037 - SUSE LINUX Enterprise Server 11: Installation notes
SAP Note 171356 - SAP Software on Linux: General information
File System
Beside the root file system, optional SAP HANA backup file systems, /usr/sap and file systems for
other applications, a special thought has to be given to /hana/data, /hana/shared and /hana/log.
30
The /hana/shared is a shared file system for transports between PRD, QAS, and TST SAP HANA
Servers, to exchange data between nodes and the location for the SAP HANA executables of a scale-
out solution. Typical options are GPFS or NFS.
Decisions made for /hana/data and /hana/log will have a direct effect on storage, network, HA and
DR solutions as well as backups11. Valid filesystem options are listed in SAP Note 2055470 - HANA on
POWER planning and installation specifics - central note .
The size of the SAP HANA file systems is determined during the sizing effort documented in the
Hardware Planning chapter. Basis for the file system sizes is the SAP HANA Administration Guide
providing detailed pictures of the file system tree and the SAP HANA Quicksizer output.
Planning Considerations for XFS with Multipathing

Using XFS with multi-pathing and LVM striping provides options to optimize the I/O characteristics.
This chapter will explain how fibre-channel connectivity, zoning and multi-pathing have an effect on
the LUN layout planning.
Note:
The minimum number of paths, volumes and disks determined in the sizing process is the absolute
minimum, even if in the following planning process less might be sufficient.
To optimize the implementation on Linux the following considerations should be taken into account
for the data and log file systems:
When increasing the number of ports, the minimum number of LUNs should be equal to the
number of active fibre-channel ports (size adjusted accordingly).
When increasing the number of LUNs, they should be a multiple of the number of FC ports.
This will later ensure to optimize the LVM striping.
The number of logical volume stripes should match the number of LUNs
Do not use the cfq scheduler (what is the default in SLES 11 SP3 and SP4).
Increase the device queue_depth to at least 64.
Quick Reference: File System Definition

Follow the SAP HANA on IBM Power Systems and IBM System Storage (Supplemental IBM
Installation Guide to the SAP HANA Master, Implementation and Administration Guides) when using
XFS or use appropriate file system documentation.
Run the verification prior to the installation of SAP HANA as described in chapter Verification.
11
Technically a single-node SAP HANA can be installed on top of GPFS. In this case it has to be ensured by the
customer that the setup still meets the SAP KPI criteria (HWCCT). A move from an XFS file-system to GPFS at a
later time is possible using SAP HANA System Replication or GPFS methods. GPFS on SAN is very sensitive in
regards to how the storage is configured and formatted down to the ranks. Please involve GPFS specialists to
create a valid setup from the disk up to GPFS and workload tuning as improper setups can impact the systems
I/O performance by a factor of 10 or more. Using IBM Spectrum Scale TDI certified systems does not erase the
requirement to verify the deployment using HWCCT. Ensure boot disk requirements and Life Partition Mobility
(LPM) support is in place as required.
31

Planning Sheets: File System

Table 10. SAP HANA File System
No Mount point Description FS size Comment

1 OS root file system / Supported by OS
2 Min 1.2x net disk space, typically
SAP HANA Data xfs for scale-up and striped over
min. 4 LUNs
3 Min. 1/2* Ram Max 512 GB,
SAP HANA Log typically xfs for scale-up and
striped over min. 4 LUNs
4 Typically NFS or GPFS. Sacle-up:
MIN(1 x RAM; 1 TB), Scale-out: 1
SAP HANA shared
x RAM_of_worker per 4 worker
nodes
5 SAP HANA FS Backup Optional. Depends on Backup
target Software
6 SAP HANA Binaries
SAP HANA Software

The SAP HANA database should be installed according to the SAP HANA Installation Guides.
Planning Considerations
To be able to move the business data to an SAP HANA deployment, verify the minimum SAP
application levels which need to be met. Please verify with SAP to get the latest minimum releases
or check the SAP PAM.
For the SAP HANA installation the hdblcm tools must be used (do not use hdbinst anymore).
For scale-out deployments consider following:

The recommendation from SAP and IBM is to first scale-up to the maximum before starting
to scale out. SAP should perform better and the overall maintenance costs will be reduced.
See SAP Note 2055547 for the amount of LPARs which can share a physical server. Fitting all
the LPARs in a physical server using VIOS benefits from advanced network throughput over
the internal v-lan.
SAP enforces to install every scale-out deployment with the Auto Host failover
independently if SAP HANA System Replication is in place or not. Verify with SAP starting
with SPS12 if this has changed.
Quick Reference: SAP HANA Software.

Please follow the below referenced SAP HANA Installation Guide and the SAP HANA on IBM Power
Systems and IBM System Storage Supplemental IBM Installation Guide.
32
Links, References, and Tools

SAP Note 218464 or 2227464
Planning Sheets: Software and OS

Table 11. SAP Software Planning Sheet
No SAP Software planning Sheet Description Notes

1 SAP HANA Release SAP Note 218464 or 2227464
2 HWCCT update SAP Note 1943937
3 SAP Application / Release SAP Note 218464 or 2227464
4 Linux SAP Note 2055470
5 SAP HANA Installation Type SAP HANA System Replication, scale-up/out, ...
Table 12. SAP HANA Installation Planning Sheet
No SAP HANA Installation planning sheet Description Notes

1 SAP HANA SID
2 SAP HANA Instance Number
3 "Planning Considerations for Virtual
Virtual IP(s) IPs and "SAP System Replication
requirements"
Verification
Before installing HANA the setup must be verified with the SAP HANA Hardware Configuration Check
Tool (HWCCT). It provides several modules to verify the configuration, but cannot guarantee a 100%
perfect setup.
For SAP HANA on POWER in a scale-up deployment, only the Landscape- and the Filesystem-Test are
relevant to validate the environment before going life or after configuration changes.
Ensure to keep ongoing records of the HWCCT documentation containing the KPIs for the file-system
test and the HWCCT version used. This is useful in case of support situations and for regular health
checks.
The HWCCT tool:
MUST be used before go life

SHOULD be used after each change in the landscape
Is RECOMMENDED for regular health checks
The minimum KPI figures and instructions about how to run and configure HWCCT can be found in
SAP Note 1943937.
33
Optional Software and Hardware Considerations

This section lists optional IBM products in a SAP HANA on POWER SAP landscape. Other vendors
might provide similar products which can be used too.
List the chosen Software in Table 13.
Table 13. Eco System Software for SAP HANA
documentation / resource
No Name Purpose
requirements /architecture prereq.
1
2
3
4
5
Migration
SAP SUM (Software Upgrade Manager) tool with the DMO option can be used to migrate from a
traditional database to the SAP HANA In-Memory database and from the traditional Suite to
S4/HANA.
SAP HANA System Replication can be used for a zero downtime migration between servers with the
same endianess. Use cases are:
From scale-up to scale-out

From XFS to IBM Spectrum Scale (GPFS) file system
Between Power Servers
SAP HANA heterogeneous system copy can be used to move the data to a new database
independent from endianess.
Starting with SPS9, HANA software additional upgrade options are available as documented by SAP.
The SAP HANA Administration Guide and SAP HANA Master Guide provide the necessary SAP
information. IBM specifics are documented in the SAP HANA on IBM Power Systems and IBM
System Storage (Supplemental IBM Installation Guide to the SAP HANA Master, Implementation
and Administration Guides)
Security
For HANA on Power similar security mechanisms are in place as for appliances. The server sided
deployment model provides more options than a typical appliance but also comes with additional
responsibility.
For third party security tools ensure the vendor provides full support for his product. To analyze
problems through SAP Incidents verify that the tool can be turned off to be able to differentiate
between SAP and vendor responsibility.
The core documentation from SAP is the SAP HANA Security An Overview whitepaper.
34
For OS hardening SUSE has published a guide specifically for SAP HANA on SUSE SLES:
https://www.suse.com/promo/security-guide-for-sap-hana.html
In case of questions on how to map the SAP Security guidelines please contact the ISICC info service
(isicc@de.ibm.com) for guidance.
Backup
SAP HANA provides several methods for data protection with their pros and cons. The major
differences between these are:
How to order, maintain, schedule and restore the backups in case needed.
Which product is used to store the backups in a secure and safe location
These methods are the same for Intel and Power. Below some products are highlighted which can be
considered as a useful extension.
IBM Spectrum Protect12

IBM Spectrum Protect provides a flexible file-system based backup method working for XFS as
well as for IBM Spectrum Scale (GPFS) file -systems. The traditional backup-archive client can be
used to make file system based backups of SAP HANA data and log as well as the required
configuration files. The source for the backup can be a SAP HANA file system backup, a SAP HANA
snapshot target or directly from the offline database.
It can be also used to cleanup growing archive logs of the database to prevent from full file -systems.
IBM Spectrum Protect (Version 4.1.0.2) is BACKINT certified also for SAP HANA.
A list of supported storage subsystems can be found here: http://www-

01.ibm.com/support/docview.wss?uid=swg21455924
SAN Infrastructure and SAP HANA File System Backup

For SAP HANA file system backup, a target file system for the backups has to be planned. When
scheduling automated backups, it will be vital to monitor for full log backup directories, as full target
file systems will suspend SAP HANA database operations. With SAN implementations, the target file
systems can be easily located to a second storage system to protect the data physically.
SAP HANA Snapshots

SAP HANA Snapshots is an option provided and supported by SAP HANA to create a fast copy of SAP
HANA production data with minimal interruption in writing the persistency backup. This option
requires a FlashCopy license for the SAP HANA data area on the storage subsystem. It has to be
ensured to transport the snaps to a backup archive location.
To perform SAP HANA snapshot functionality, check the IBM FlashCopy backup solution for SAP
HANA TDI documentation. It describes the required setup and outlines the process including scripts
to reuse on an SVC based sample setup. A list of supported storage subsystems can be found here:
http://www-01.ibm.com/support/docview.wss?uid=swg21455924
12
Formerly: Tivoli Storage Manager (TSM)
35
IBM Spectrum Scale and Elastic Storage Server

GPFS8 has native snapshot and encryption functionality which can be exploited performing SAP
HANA Snapshots. Using IBM Spectrum Scale ensure to run on a stack which supports Life Partition
Mobility (LPM) in case this is an operational requirement. Using Elastic Storage Servers verify options
available for boot disks.
Verify SAP Note 2055470 for file system support.
High Availability and Disaster Recovery

All native SAP HANA High Availability (HA) and Disaster Recovery (DR) methods can be used for SAP
HANA on POWER deployments as documented by SAP.
The following SAP Documentation gives a good overview: Introduction to High Availability for SAP
HANA
SAP HANA System Replication (Multi-tier)

The key functionality in SAP HANA for High Availability and/or Disaster Recovery is SAP HANA System
Replication. This is also available in a multi-tier version, providing the option to have both a near
standby and a remote standby.
For pure DR a manual takeover might be sufficient. Automated failure detection and takeover
procedures are required to minimize the operational costs and the takeover time (RTO) for HA
purposes.
IBM Tivoli System Automation Multi-Platform
SA MP can be used to cluster SAP HANA System Replication instances. It also provides products to
automate the HA setup for the required application servers, SAP HANA and other components of a
SAP Landscape out of the box. Looking towards enhanced options, SA MP is GPFS aware and can
support up to 128 nodes for scale-out deployments. For DS8000 a HyperSwap deployment can be
considered for native storage attachments using SA MP.
The policies are included in 4.1.0.2. Please go to following resources for solution and documentation
download:
http://www-01.ibm.com/support/docview.wss?uid=swg24039033
http://www.ibm.com/support/knowledgecenter/SSRM2X_4.1.0.2/com.ibm.samp.doc_4.1.0.
2/sampicpart4sapha_HANA.html
SAP HANA Auto Restart, Auto Host failover, Failover Hooks

SAP HANA supports out of the box HA capabilities. They can be used for both scale-up and scale-
out. This requires appropriate SAN zoning along with the shared disc installation or GPFS along with
the shared filesystem installation.
IBM Storage Replication

The IBM Storage Subsystems can be used to mirror SAP HANA data to a DR site using consistency
groups. Synchronous and asynchronous methods can be applied.
36
Management and Monitoring

PowerVC
IBM PowerVC Virtualization Center is an advanced virtualization management offering, built on
OpenStack, that provides simplified virtualization management for IBM AIX, IBM i and Linux
running on IBM Power Systems.
PowerVC is designed to improve administrator productivity and simplify the management of VMs
and LPARs on Power Systems servers. PowerVC provides the foundation for Power Systems scalable
cloud management, including integration to higher-level cloud managers based on OpenStack
technology.
PowerVC can be used to rapidly deploy partitions including zoning and disk provisioning. It builds a
good match to other OpenStack components and is pure Power focused and by that well integrated
into the technology options.
Support and Services
Getting help and information from the World Wide Web for IBM products
On the World Wide Web, the IBM website has up-to-date information about IBM systems, optional
devices, services, and support. You can find service information for IBM systems and optional
devices at http://www.ibm.com/supportportal.
Getting help and information for IBM Server and Storage mapping
People responsible for SAP HANA sizing who are not familiar with the SAP process and have not
received quantified system requirements (memory GB, cores or SAPS) for the SAP HANA project can
contact their responsible IBM TechLine, FTSS, or ATS team for guidance through the official SAP
sizing process.
Standard Support Flow

Not using the centralized IBM service, support and operation offerings for SAP HANA it has to be
verified the product owner provides sufficient support. Figure 10 outlines the standard support
ownership of SAP HANA, the operating system and the hardware stack. Figure 11 brings the typical
responsibilities into the context of SAP Support.
Note:
To collect OS and HANA system configuration information use the sapsysinfo and HWCCT tool
available through following SAP Notes:
SAP Note 618104 - "sapsysinfo" - Compiling system information on Linux
SAP Note 1943937 - Hardware Configuration Check Tool - Central Note
37
Component support
Component
provider:
owner:
SAP IBM Vendor
SAP AG SAP HANA
SAP AG HWCCT - TOOL
OS Vendor Linux Operating System
IBM IBM Power Server and PowerVM
Product Vendor SAN IB Ethernet
IBM SVC (optional) GPFS
(IBM) Storage
Product Vendor
Server
Figure 10. Standard Support ownership of SAP HANA and Hardware Stack
Customer faces an issue with HANA on

Power Servers
no Customer can
Open SAP OSS
identify the origin
Ticket on HAN-*
area of the problem
yes
Open SAP OSS
Ticket on HAN-*/ yes SAP code/ OS
BC-OP-PLNX/ configuration/OS
BC-OP-LNX/ related
Issue within no Origin area

SAP
identified
responsibility
Storage/ yes Customer opens a
yes Infrastructure problem record at the
related vendor
Resolution remains in no
regular SAP Support
Process Server yes Customer opens an
related IBM PMR
no
Open SAP OSS

Ticket on HAN-*
Customer opens a problem yes rootcause in no Resolution remains in

record at SAP referencing SAP code/ regular partner support
the PMR/Incident number. config/OS process
Figure 11. Support Flow for HANA on Power Solution Stack
IBM Services, Support and Operation offerings for SAP HANA on Power
Besides the SAP support and standard IBM support for the infrastructure components, IBM intends
to provide at GA a more comprehensive set of services to best support customers for SAP HANA on
Power systems.
The following areas will be addressed by the services
38
Planning and Installation

The planning process introduced in this paper which is based on a SAP sizing and an IBM mapping
supported by the IBM TechLine Enablement (SAP Sizing Questionnaire for Suite and BW) can be
enriched by the IBM Systems Lab Services. These services will offer clients services for on-site SAP
HANA Installations.
Integrated Support
TSS enhances the TSS Center of Competence for SAP HANA on POWER providing specialists for
SAP HANA infrastructure support.
The CoC for SAP HANA on POWER provides a comprehensive set of services building a solution
support with Single Point of Contact (SPoC), integrated HW/SW support for the infrastructure and
proactive services. The offering helps customers to ease management and improve uptime with a
single source of integrated break-fix, how-to and proactive support for hardware and software.
The Total Support for SAP HANA on Power Systems consists of two major services:
Solution Support Base Module for SAP HANA on Power Systems
This is an integrated solution support with Single Point of Contact, integrated HW/SW break fix,
how to and coordination of all proactive services.
Proactive Services
A set of proactive remote services like Proactive System Check for SAP HANA on Power,
Microcode and Release Management and an onsite Health Check will help customers to identify
inconsistencies early or avoid problems.
Please contact your local TSS representative.
Proactive services can provide:
System analysis to identify inconsistencies

Assistance to keep your systems up to date Clients
Remote and onsite support
Single-source support
Optimized reactive services can provide:
Integrated support for Power Systems Optional project services by

infrastructure (hardware and software) IBM or IBM Business Partner
Faster response time for mission-critical

Design and strategy
environments
Implementation, disaster
Access to the rich resources in our Center
recovery and backup
of Excellence
Configuration chances and
Customer care, including quarterly meetings
patch management
with your IBM focal
Figure 12. IBM Services for SAP HANA on POWER environments
39
Referenced documents
This is a full list of documents used to write this planning Guide. Some of these require special
permissions. Collaborating with IBM, SAP or your Business Partner will provide access to all of them.
SAP Notes:
SAP Note 618104 - "sapsysinfo" - Compiling system information on Linux

SAP Note 1310037 - SUSE LINUX Enterprise Server 11: Installation notes
SAP Note 1514966 - SAP HANA 1.0: Sizing SAP In-Memory Database
SAP Note 171356 - SAP Software on Linux: General information
SAP Note 1736976 - Sizing Report for BW on HANA (SAP Note 1637145 outdated)
SAP Note 1813548 - Database Migration Option (DMO) of SUM
SAP Note 1872170 - Suite on HANA sizing report (earlier SAP Note 1793345 )
SAP Note 1943937 - Hardware Configuration Check Tool - Central Note
SAP Note 1954788 - SAP HANA DB: Recommended OS settings for SLES 11 / SLES for SAP
Applications 11 SP3
SAP Note 1944799 - SAP HANA Guidelines for SLES Operating System Installation
SAP Note 2055470 - HANA on POWER planning and installation specifics - central note
SAP Note 2133369 - SAP HANA on Power: Central Release Note (Valid for SPS9 and SPS10, it
is replaced for SPS11 by SAP note 2227464)
SAP Note 2161344 - HWCCT patch note
SAP Note 1900823 Storage Connector API (for Auto Host Failover deployments)
SAP Note 2230704 - SAP HANA on IBM Power Systems with multiple - LPARs per physical
host
SAP Note 2188482 - SAP HANA on IBM Power Systems: Allowed Hardware
SAP Note 1953429 - SAP HANA and SAP NetWeaver AS ABAP on one Server
SAP Note 218464 - Supported products when running SAP HANA on IBM Power Systems
SAP Note 2227464 - SAP HANA Platform SPS 11 Release Note
SAP Documents:
Sizing:
o SAP HANA sizing (SAP Community Network, Registered S-Users)
o Sizing Approaches for SAP HANA Lessons Learned
o How to size SAP BW on SAP HANA
o Sizing for SAP Business Suite powered by SAP HANA
o Quick Sizer Tool (ensure to use the HANA Quick Sizer)
Storage and High Availability

o SAP HANA Storage Requirements
o SAP Certified Enterprise Storage Hardware for SAP HANA
o SAP HANA Network Requirements
o Introduction to High Availability for SAP HANA
Architecture
o SAP HANA Platform (all HANA related documents as in scope accordingly to SAP
Note 2130682)
40

SAP HANA Administration Guide
SAP HANA Master Guide
SAP HANA Security Guide

o SAP HANA core documentation sets: http://help.sap.com/hana
o SAP HANA in Data Centers

o IT Planning Documents
o SUSE LINUX Enterprise Server Deployment Guide
o SUSE Knowledge Base
o SAP HANA TDI FAQ
IBM Documents:
Assistance
o IBM TechLine support
o SAP Sizing Questionnaire for Suite and BW
o ISICC SAP HANA on Power Systems (IBM only)
Storage
o IBM System Storage Architecture and Configuration Guide for SAP HANA TDI
o SAPmagic - Storage Sizing Tool for SAP landscapes (IBM and BP only)
o IBM FlashCopy backup solution for SAP HANA TDI
Hardware mapping:
o ISICC SAPS Capacity Tables for Power Systems (IBM only, BPs and customers
should contact the IBM TechLine for support)
o POWER8 Facts and features
o POWER7+ Facts and features
o Sales Support Information (SSI) (IBM only)
Deployment
o SAP HANA on IBM Power Systems and IBM System Storage (Supplemental IBM
Installation Guide to the SAP HANA Master, Implementation and Administration
Guides)
o IBM FlashCopy backup solution for SAP HANA TDI
41
Copyrights and Trademarks

Copyright 2015 IBM Corporation. All Rights Reserved.
IBM Corporation
Systems and Technology Group
Route 100
Somers, NY 10589
Produced in the United States of America
Neither this documentation nor any part of it may be copied or reproduced in any form or by any
means or translated into another language, without the prior consent of the IBM Corporation.
IBM makes no warranties or representations with respect to the content hereof and specifically
disclaim any implied warranties of merchantability or fitness for any particular purpose. IBM
assumes no responsibility for any errors that may appear in this document. The information
contained in this document is subject to change without any notice. IBM reserves the right to make
any such changes without obligation to notify any person of such revision or changes. IBM makes no
commitment to keep the information contained herein up to date.
Edition Notice: 2016

This is version 2.4 of this document.
IBM, the IBM logo, and ibm.com are trademarks or registered trademarks of International Business
Machines Corp., registered in many jurisdictions worldwide. A current list of IBM trademarks is
available on the web at "Copyright and trademark information" at:
http://www.ibm.com/legal/copytrade.shtml.
Adobe and PostScript are either registered trademarks or trademarks of Adobe Systems
Incorporated in the United States and/or other countries.
Intel, Intel Xeon, Itanium, and Pentium are trademarks or registered trademarks of Intel Corporation
or its subsidiaries in the United States and other countries.
Java and all Java-based trademarks and logos are trademarks or registered trademarks of Sun
Microsystems, Inc., in the United States, other countries, or both.
UNIX is a registered trademark of The Open Group in the United States and other countries.
Linux is the registered trademark of Linus Torvalds in the United States, and other countries.
Microsoft, Windows, Windows NT, and the Windows logo are trademarks of
Microsoft Corporation in the United States, other countries, or both.
SAP HANA, SAP NetWeaver are trademarks or registered trademarks of SAP Corporation in the
United States, other countries, or both.
Oracle is a registered trademark of Oracle Corporation and/or its affiliates.
Other company, product or service names may be trademarks or service marks of others.
Information is provided "AS IS" without warranty of any kind.
Information concerning non-IBM products was obtained from a supplier of these products,
published announcement material, or other publicly available sources and does not constitute an
endorsement of such products by IBM. Sources for non-IBM list prices and performance numbers
are taken from publicly available information, including vendor announcements and vendor
worldwide home pages. IBM has not tested these products and cannot confirm the accuracy of
performance, capability, or any other claims related to non-IBM products. Questions on the
capability of non-IBM products should be addressed to the supplier of those products.
42
Disclaimer and Special Notices

This information could include technical inaccuracies or typographical errors. Changes are
periodically made to the information herein; these changes will be incorporated in new editions of
the publication. IBM may make improvements and/or changes in the product(s) and/or the
program(s) described in this publication at any time without notice.
Any references in this information to non-IBM Web sites are provided for convenience only and do
not in any manner serve as an endorsement of those Web sites. The materials at those Web sites are
not part of the materials for this IBM product and use of those Web sites is at your own risk.
Any performance data contained herein was determined in a controlled environment. Therefore, the
results obtained in other operating environments may vary significantly. Some measurements may
have been made on development-level systems and there is no guarantee that these measurements
will be the same on generally available systems.
Furthermore, some measurement may have been estimated through extrapolation. Actual results
may vary. Users of this document should verify the applicable data for their specific environment.
Information concerning non-IBM products was obtained from the suppliers of those products, their
published announcements or other publicly available sources. IBM has not tested those products
and cannot confirm the accuracy of performance, compatibility or any other claims related to non-
IBM products. Questions on the capabilities of non-IBM products should be addressed to the
suppliers of those products.
This information contains examples of data and reports used in daily business operations. To
illustrate them as completely as possible, the examples include the names of individuals, companies,
brands, and products.
All of these names are fictitious and any similarity to the names and addresses used by an actual
business enterprise is entirely coincidental.
COPYRIGHT LICENSE:
This information contains sample application programs in source language, which illustrate
programming techniques on various operating platforms. You may copy, modify, and distribute
these sample programs in any form without payment to IBM, for the purposes of developing, using,
marketing or distributing application programs conforming to the application programming interface
for the operating platform for which the sample programs are written. These examples have not
been thoroughly tested under all conditions. IBM, therefore, cannot guarantee or imply reliability,
serviceability, or function of these programs
ANY INFORMATION HEREIN IS PROVIDED AS IS WITHOUT WARRANTY OR INDEMNIFICATION OF

ANY KIND BY IBM AND DO NOT ANY EXPRESS OR IMPLIED, WARRANTIES OR CONDITIONS OF
MERCHANTABILITY, FITNESS OR USAGE FOR PARTICULAR PURPOSE AND ANY WARRANTY OR
CONDITION OF NON-INFRINGEMENT.

SAP HANA On Power-Planning - 3.0

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

SAP HANA On Power-Planning - 3.0

Uploaded by

Copyright:

Available Formats

SAP HANA on

IBM Power Systems

IBM Deutschland Research & Development GmbH

Edition Notice and Version Information

Produced in the United States of America

Edition Notice: 2016

Doc Version Changes

About This Document

Planning Considerations Multi-Host SAP HANA (scale-out) ......................................................... 26

Before starting, please download or check

SAP Note 2130682

SAP HANA on POWER Solution Overview

Figure 1. SAP HANA on POWER Solution Stack

Figure 2. Supported Application Server Environments

Obtain required hardware capacity

Complete SAP HANA Apply SAP tools versus

Identify Power System with sufficient DIMM and

Calculate socket/core count

Enforced by # of DIMMs Match sizing output with SAPS capacity

Identify and Layout Storage System(s)

Meet RTO and RPO Match sizing output with

Map LAN & SAN I/O adapters into

Verify SAP KPI-compliance of the

HANA System deployment

Figure 3. SAP HANA on IBM Power Servers Planning Process

Mapping SAP Sizing Output to IBM Power Systems Configurations

SAP Sizing for SAP HANA on POWER

Two basic SAP HANA sizing methods can be performed:

Source SAP BW Table Sizes

Target SAP HANA DB Size

For scale-up these values

More Details (growth etc.)

Figure 4. Sample get_size report output

Memory and CPU Capacity background information

Background information on core to memory ratio

Disk Quantification Background information

Quick Reference: Get SAP HANA Sizing from SAP

Need yes contact your respective TechLine, FTSS

Business yes Apply SAP tools versus existing DB

Complete SAP Quicksizer process or use the SAP Sizing

Quantified system requirements (memory GB, cores, SAPS, volume)

Figure 5. Quick Reference: "Get SAP HANA Sizing from SAP"

Links, References and Tools

Central SAP Notes about SAP HANA sizing:

Planning Sheets: SAP Sizing

No SAP Planning Sheet SAP Sizing Result Notes

Mapping SAP Sizing Output to IBM Power Systems Configurations

For scale-out (multi-host) the inter-LPAR ethernet communication, the auto-host-failover

The size of the SAP HANA partition

Planning Considerations for Power Systems Memory Capacity

Planning Considerations for Power Systems CPU Capacity

Virtualization / LPARs performance related considerations

Planning Considerations for Power Systems I/O Adapter Capacity

Quick Reference: Find valid IBM Power Systems options

is this server for

Add additional requirements

List of valid IBM Power Servers available.

Sandbox and non-production Floor configuration eligable to Ceiling configuration eligable to

Links, References and Tools

POWER8 Facts and Features

Sales Support Information (IBM only)

Planning Sheets: HANA Partition and IBM Power Server

No Partition Planning Sheet # per HANA partition Sum Notes