1. Migration from 8.1 to 11.3
NLS Consideration:If NLSisenabledonthe source system, we shouldbe sure thatNLSis
enabledonthe destinationaswell.Alsothe defaultconfigurationneedtobe checkedif theyare
similar.If the configurationisnot same inboth,needtochange it tocustom.
There are newfunctionswhichare notpresentin8.1 i.e.Loopingtransformer,savedinput
record,lastrowingroup(),
The DataStage componentbackupshouldbe takencarefullyasthe driverswill change
The DSParams parameterfile shouldbe checkedandconfiguredcarefully
The uvodbc.config,odbc.ini,brandedodbcfilesshouldbe checkedforthe driverversions
References:
https://www-
01.ibm.com/support/knowledgecenter/#!/SSZJPZ_11.3.0/com.ibm.swg.im.iis.productization.iisi
nfsv.whatsnew.doc/topics/whats_new_11.3.html
http://www-
01.ibm.com/support/knowledgecenter/#!/SSZJPZ_11.3.0/com.ibm.swg.im.iis.productization.iisi
nfsv.migrate.doc/topics/top_of_map.html
http://www-
01.ibm.com/support/knowledgecenter/#!/SSZJPZ_11.3.0/com.ibm.swg.im.iis.productization.iisi
nfsv.migrate.doc/topics/sequence.html
http://www-
01.ibm.com/support/knowledgecenter/SSZJPZ_11.3.0/com.ibm.swg.im.iis.productization.iisinfs
v.migrate.doc/topics/a_performing_manual_migration_parent.html?lang=en
Order of Migration
The followingprocedures are the guide forthe migration:
Migratingcredentials
MigratingInfoSphere DataStage
MigratingInfoSphere QualityStage
Migratingcommonmetadata
MigratingIBM InfoSphere DataQualityConsole
MigratingIBM WebSphere RTI
MigratingInfoSphere BusinessGlossary
MigratingInfoSphere FastTrack
MigratingInfoSphere InformationAnalyzer
MigratingInfoSphere MetadataWorkbench
MigratingInfoSphere InformationServicesDirector
Migratingreports
2. Migrating Credentials:
Credentialsmigrationisnotsupported:PriortoVersion8.5,the exportof credentialsisnot
supported.Youwill needtomanuallyrecreatethe credentialsusingthe InfoSphere®
InformationServerWebconsole.
Migrating Datastage
Migrating from InfoSphere DataStage
Complete thesetaskstomigrate IBM® InfoSphere® DataStage®.
About this task
Before youuse thisprocessto migrate jobs,reviewthe jobsthatyouplanto migrate to determine which
itemsmightrequire manual intervention.The followinglistdescribesthe additional itemsthatyou
mightneedtomanuallymove tothe target:
DSParamsfor each projectandthe DSParams inthe template fornew projects
User-modified IBMInfoSphere QualityStage® overrides
The FTP/Sendmail templateinthe projectdirectory
The uvodbc.configfileinthe projectdirectory
MessageHandlers,whichare under the Serverdirectory
Jobcontrol language (JCL) templates(DS390)
Parallel engine configuration uvconfigfilewhichcontainsspecificoptionsforthe environment
User-definedentriesinthe dsenv file
Data sourcesin the odbc.ini file
Parallel engine maps andlocales
Parallel engine configurationfiles
The followinglistdescribesthe additional tasksthat we mightneedtoperform:
Recreate usernamesandcredential mappings
Run the ConnectorMigrationtool to update connectors
If you are migratingfrom InfoSphere DataStage,Version7.5andearlier,andhave jobsthat are using
ORAOCI8plugins,thenyoumustfirstconvertthe jobsto use the ORAOCIplugin,whichcansupport
eitherOracle 8 or Oracle 9. Youmust convertthe jobsbefore exportingthemtothe source computer.
3. Exporting InfoSphere DataStage projects
Complete thesetaskstoexport InfoSphere® DataStage® projects.
1. Capturingjobloginformation
If you planto remove the source installationandreplace itwiththe targetinstallation,save the
joblog information,whichincludesenvironmentsettingsandotherinformationthatyoulater
use to validate the resultsonthe targetsystem.
2. Backingup the installation
Before youbeginexporting InfoSphere InformationServer fromyoursource computerandafter
youcomplete the importprocessonthe targetcomputer,youshouldback upthe installation.
3. SavingInfoSphere DataStage settingsfiles
Save the settingsfilesfromthe source installation.Thenafteryouinstallthe new version,
integrate the savedsettingsintothe settingsfilesonthe targetinstallation.
4. Movingjobdependencyfiles,hashedfiles,andjoblevelmessagehandlers
If the jobsinthe source installationdependonfilessuchasflatfiles,schemafiles,libraryfiles,
and hashedfilesthatare locatedindirectorystructuresthatwill notbe accessible fromthe
target installationof InfoSphereDataStage,youmustsave the filesandmanuallymovethemto
the target installation.
5. Exportingthe projects
Use the istool commandline interface (CLI) toexportall versionsof InfoSphere
DataStage projects.Youcan alsouse the dscmdexportcommandorthe InfoSphere
DataStage Manager clientforVersion7.5.3 or earlier.Youcan use the InfoSphere
DataStage DesignerclientforVersion8.0.1or laterto exportInfoSphere DataStage projects.You
can use the InfoSphereInformationServer ManagerforVersion8.1 or laterto export InfoSphere
DataStage projects.If youuse the istool ordscmdexportcommand,youcancreate a scriptthat
exportsall projectsatone time.
Removingthe InfoSphere DataStage serverand clients
If you planto replace the existingversionwiththe new version,remove the InfoSphere
DataStage serverandclientsbefore youinstall the new version.
Installingthe newversionof InfoSphere InformationServeron the clientcomputer
You do notmigrate the client;instead,youinstallthe new versionof the clientprogramsonthe
clienttier.
Importing InfoSphere DataStage projects
Complete thesetaskstoimportIBMInfoSphere DataStage projectsintothe new version
of InfoSphereInformationServer.
1. Importingprojectinformation
4. Use the istool commandline interface (CLI) toimport InfoSphere DataStage projects,for
Versions8.5and later.You can alsouse the dscmdimportcommand,the InfoSphere
DataStage Designerclient,orthe InfoSphere InformationServer Mangerclienttoimport
projects.
2. Mergingthe contentsof the InfoSphere DataStage settingsfiles
For Version8.1and earlier,use the Administratorclient ordsadmincommandline tomanually
create the environmentvariablesinthe new system.ForVersion8.0.1andearlier,use the
Administratorclienttoperformthese tasks.
3. Restoringjobdependencyfilesandhashedfiles
Restore jobdependencyandhashedfilestothe new installation.
4. Recompilingjobs
Before youcan run jobsand routines,youmustrecompile them.
5. New Features in 11.3
Introduction
With the release of Information Server/DataStage 11.3 a few weeks ago, most DataStage developers are
interested in knowing exactly what new features have surfaced and how they can best be leveraged. With
the release of version 8.7, IBM introduced the Operations Console and version 9.1 followed in-line with
the release of the Workload Manager. I’m afraid that DataStage developers don’t have anything too
exciting to look forward to in version 11.3. There are definitely some nifty new features tacked on the suite
from the standpoint of data governance, metadata management, and administration, but this post will
review just the new features in DataStage.
There might be some hidden new features or “features” which aren’t documented. Feel free to comment
below on what you think they might be.
Hierarchical Data Stage
Remember how the XML stage was pretty recently introduced for all XML processing in DataStage? Well
now it has been relabeled as the Hierarchical Data stage, I suppose to account for its ability to process all
types of Hierarchical Data (JSON) as opposed to strictly being limited to XML. This stage also has some
additional functionality which wasn’t previously available. If you are familiar with this stage (Hierarchical
Data/XML) you will know it has various steps which are added in the Assembly Editor, for a sequence of
processing events. There are now three new steps:
REST – Invokes a RESTful web service
JSON_Parser – Parse JSON content with a selected type
JSON_Composer – Compose JSON content with a selected type
6. Big Data File Stage
The Big Data File stage is used to read and write to files on Hadoop (HDFS). The Big Data File stage is
now compatible with Hortonworks 2.1, Cloudera 4.5, and InfoSphere BigInsights 3.0.
Greenplum Connector Stage
You can now use the Greenplum Connector stage for a native connection for accessing data which is
located in a Greenplum database. You can now also import Table Definitions using the Greenplum
Connector framework.
InfoSphere Master Data Management Connector Stage
The Master Data Management Connector stage can be used to read and write data from the IBM master
data management solution – InfoSphere MDM. This stage can be configured for Member read and
Member write interactions from the MDM server.
Amazon S3 Connector Stage
Amazon S3 (Simple Storage Service) is a cheap cloud file storage system which offers availability
through web services (REST, SOAP, and BitTorrent). It offers scalability, high availability, and low latency
at extremely competitive prices. The Amazon S3 Connector stage be can used to read and write data
residing in Amazon S3.
7. Unstructured Data Stage – Microsoft Excel (.xls and .xlsx)
The Unstructured Data stage was first introduced in DataStage v9.1 and was used to read Excel files
through a native interface. Previously, Excel data was staged as a .csv file or accessed through ODBC.
The stage can also now be used to write data to Excel files.
Sort Stage Optimization
The Sort stage now tries to optimize your DataStage sort operations by converting length bounded
columns to variable length before the sort and then converts it back to a length bounded column after the
sort. When a record’s actual size of data is smaller than the defined upper bound, the optimization will
result in reduced disk I/O.
Improved Flexibility in Record Delimiting
The Sequential File stage now gives developers more flexibility with how a source flat file has to be
delimited. A new environment variable, APT_IMPORT_HANDLE_SHORT, can be set to enable the
import operator the ability the read in records which do not contain all of the fields defined in the import
schema. Previously, these records were rejected by the stage. The values assigned to any missing field
depends on the data type and nullability.
Operations Console/Workload Management
IBM lists the Operations Console and Workload Management as new features of the 11.3 release
documentation, even though these components have already been introduced in previous releases. Both
components are now part of the base Information Server installation and Workload Management is now
by default enabled.