Documentos de Académico
Documentos de Profesional
Documentos de Cultura
December 2013
Agenda
Targeted audience:
Technical Architect
BASIS/Netweaver consultant
Hardware Software
Multi-Core Architecture
+
(8 x 10core CPU per blade)
Parallel scaling across Row and Compression
Column Store
blades
SAP Hana
SAP Hana ui for SAP Business Apps
SAP NW 7.3
Information Suite
BW
Access
Platform/Runtime e
SAP business Primary persistence
Object BI Suite
SAP HANA
Side-by-Side
Host Agent/ R runtime : for direct
SAP HANA Database SAP Hana LM integration with R-
SMD Agent
runtime libraries
Process Description
Name server Maintains Landscape Information
Index server Holds data and execute all operations
Statistics server Collects performance Data about Hana , Pro-active alerting
XS server XS Service - application server
SAP hana studio repository Repository for HANA content
SAP Host Agent Enables remote start/stop and integration so solution manager
SAP Hana Lifecycle Manager Manage SW updates for HANA
Copyright 2012 Accenture All rights reserved. 7
Installation
Overview of the standalone SAP HANA Installation
SAP HANA Installation process: Appliance is installed and OS-configured by a HW manufacturer
certified by SAP (IBM, HP, etc.).
Activities include:
HW installation,
OS installation (SLES) and Network configuration,
HANA Installation starting from SP6 customer can install HANA itself (by a certified administrator - who has assisted
SAP HANA200 Training and pass related certification)
Tools provided:
Installation with the SAP HANA Unified Installer (based on configuration file)
Script provided by SAP to validate the installation
Some manufactures provide scripts to gather all logs / files required for OSS ticket (for example: IBM)
Components installed:
SAP HANA Platform Edition (HANA DB, client, studio, host agent, etc)
Extra functionalities can been provided by AFL, LC Live Cache APPS, etc.
Installation of additional software is not supported by SAP (unless certified): for example: backup agent, kerberos
module, etc.
Post installations:
SAP Remote support connection
Install /Check License (License is based on the Memory peak usage - for production environment)
Requirements:
About 90% of physical memory dedicated to HANA
File systems (shared, local)
HANA logfiles needs high performance disk (as opposite to datalogs)
All required disks are provided by certified HW.
Configurations possible:
Stand alone architecture
One appliance single node
One appliance multiple nodes (DEV and TST instances on the same box) Only one
customer in France (no worldwide credentials known)
Installation of additional nodes simplified with HLM (HANA Lifecycle Manager)
SAP HANA Sizing is mainly described in OSS notes for each usage:
Dedicated OSS notes for each case:
General sizing - 1514966
BW on HANA - 1736976
Business Suite on HANA - 1872170
SAP QuickSizer
Main point of concern is the definition of RAM (CPU is implicitly defined by HW)
RAM V * 2 / 7 * c.
V = Volume of data footprint
c = compression factor in source DB
/ 7 compression factor to HANA
* 2 include static and dynamic data
CPU 300 SAPS / active user (or 0.2 core / active user)
Administration and Development client of HANA, based on Eclipse (Java IDE), inherits
its ergonomic and common features.
Example of activities
Manage licence (install, update, check)
Start/Stop HANA (single, multi nodes)
Configuration (update configuration *.ini files update on files is not recommended)
Recurring administration/monitoring tasks (check status, alerts, logs, performance, etc.)
Some activities can also be performed by DBA Cockpit transaction of any SAP instance
with a DB connection to HANA box
Known issue: warning of service start sapstarsrv (IBM provided appliance)
2
4 Start import in TMS to process
SAP HANA STUDIO the file
Promotion of Data Model 5 Start import via SAP Hana
studio
HANA
TD1 TT1
1
Hana soucre Hana target
Collect data
Query Analysis :
Analyzing SQL Execution with the Plan Visualizer (graphical view)
Solution Manager
Advice : SAP Solution Manager 7.1 with SP05 : Queues in SLT can be monitored, alerts are
generated for Mainly Database Availability, Host Usage (CPU/Monitoring)
Requirements for HANA integration :
SAP Hostagent 7.20 SP 84 ( Note 1625203) + Solman 7.1 SP03
SAP diagnostics Agent has to be installed in SAP HANA box
Operations available:
Create users
Lock / unlock users
Activate / deactivate users
Reset password
Check user privileges
Assign roles / privileges
Privilege Object
5 types of privileges exist:
Type Description Purpose
Object Authorize access to data and operations on database objects Administrative tasks
System Authorize execution of administrative actions for the entire Restrict access/modifications to users
HANA DB
Application Authorize access to the HANA XS application functions Restrict access/use of application
functions of XS engine to users
Analytic Authorize read access to analytic views at runtime and Restrict access/modifications to users
provide row-level access control based on the dimensions of in the context of the view or table
the relevant view (criteria for filtering)
Package Authorize access in the repository (modelling environment) at Restrict access/use of packages to
design time users
Copyright 2012 Accenture All rights reserved. 20
User & roles management (3/3)
Miscellaneous remarks:
Privileges management works on grant philosophy. It is not possible to explicitly deny
privilege management
Password policy is based on certain rules (pwd length, complexity, blacklist terms,)
Predefined and templates roles and privileges are provided in as part of the installation.
Integration with NetWeaver Identity Management is possible since IDM 7.2 SP3 with IDM
connector: the scope of features are:
Creating/deleting users
Assigning roles
Setting passwords
IDM does not manage SAP HANA roles & privileges
Overview mechanism
Log writing and savepoints protect Data (Memory)
from power failures or system crash. SQL data, undo log information, modeling data,
Restore constraints:
A backup can be restored to a
External backups
system with the same number of Backup with 3rd party applications
nodes as the source system implementing BackInt for SAP HANA API
Backup with DBA Cookpit from NW
system if HANA connector is configured.
Copyright 2012 Accenture All rights reserved. 22
Backup / Restore Data backup (2/4)
Remark: Log
Segment
Log
Segment
Volume for Log volume = volume of HANA main
memory
Log
Log segment files are stored in an unique directory Segment
configured in global.ini file Log volume
SAP provides script in note 1651055 to perform backup of data, log and
configuration files
Instance replication
Storage replication: continuous (mirroring) between primary and backup storage over network,
synchronous or asynchronous
System replication: continuous update of the secondary system by primary system
Each node has its own name and index server (as well as its own data and log backup
files)
Statistics server is only on active server
Remark:
Management of additional host can be performed in SAP HANA lifecycle manager (in
the STUDIO)
Scale out for HA implies the creation of a standby host which is started as soon
as an active node fails.
Failure scenarios:
Blade failure: standby host (hosted on a distinct blade) takes over the function
Name server failure: another of the remaining name servers become the master name server
Index server failure (see picture below):
Master name server detects the failure and executes the failover (new server reads indexes from the shared
storage)
Data located in memory that is not saved in disk
Server 2 can be manually added back into the cluster through commits or save points cyclic is lost.
X
Index Server Index Server
Shared storage
The name server reads indexes from the shared
storage and the existing connections from Server 2
System replication: each primary system has a exact copy as standby system
(shadow instance).
Option1: Data and logs are continuously transferred to the secondary site with preloaded data (fast
switch over)
Option2: Data is only initially transferred, then continuous log transfer & log replay on secondary site
(even faster switch over - < 5 min).
The secondary site has preloaded data with lasted commits
Both options imply manual switchover (decision should be based on monitoring tools (refer HANA Administration
Guide)
Storage replication
Available for scale
out architecture Clients Applications
SAP supported
mirroring solutions
Data center 1 Data center 2
for storage may vary
upon vendor: OS: DNS, hostnames,
For example, IBM uses
GPFS instead of disk
arrays (even a single Primary - active Secondary
IBM HANA instance standby
uses one GPFS node) Index Server Index Server
OSS Note 1755396 for Master name Server Master name Server
details
System replication
Clients Applications
Near Zero Downtime Upgrade is possible for HANA with the system replication since
secondary system can have higher software version than the primary one.
Audit can be enabled and disabled for the entire database only
Audit Policies define which actions in the database are logged
Events can be audited Details
Change to user authorization Create/drop user, grant/revoke role
Authentication of users Connection attempts of users to database
Change to system configuration Change to ini.file, Install license key
Change to auditing configuration Tables, Views, procedures with actions like
select, update, execute..
General rule:
No installation must be performed outside HW manufacturer installation, SAP SP & patchs.
Backup of HANA data via the BackInt API with 3rd party applications:
2 backups applications certified so far (updated list of OSS note: #)
Nagios (for ex.) with JDBC connection (installation of HANA Client on Nagios server)
OSS notes
OSS# Description
1755396 Solutions for SAP HANA System replication
1651055 Backup script to perform backup of data, log and configuration files
Useful tables