IBM have a version of Planning Analytics on AWS (and soon to be Azure) called Planning Analytics as a Service (PAaaS). It uses Cloud Pak for Data on top of Red Hat Openshift to deliver a multi tenanted, high performance environment. Planning Analytics is then embedded as a container inside Cloud Pak. This post takes you through the process required to migrate from on premise TM1 to Planning Analytics as a Service.
Objects that Cannot be Migrated
These object cannot be migrated from on premise TM1 to PAaaS:
- Sandboxes
- Planning Analytics objects whose name is not a valid UTD-8 encoded string
- Cubes, views, dimensions, and subsets with names containing invalid characters. Invalid characters are
\/:*?"<>|';, - Data files in client directories
Things to Do Prior to Migrating
Before you migrate to Planning Analytics as a Service, you need to prepare. In addition to ensuring the objects identified above are not included in the migration, you need to:
- Have free disk space at least 2.5 times the size of the original version 11 database.
- Have enough RAM to load it on the machine used to run the migration.
- Perform SaveDataAll to make sure all in-memory changes materialized to their cube files.
- Make sure that the database’s Tm1s.cfg configuration file exists and it refers to the correct database directory. The configuration file must contain a DatabaseDirectory=<path> line, where <path> is the path to the actual data files relative to the Tm1s.cfg file.
- Delete or repair views that refer to elements that no longer exist. If a dimension with multiple hierarchies has the same leaf element defined with conflicting element types in different hierarchies. The migration tool will stop running when it first encounters a dimension where a conflicts exists.
Migrating Planning Analytics Local Client
One common question that will arise when migrating to CloudPak is, what will happen to the private views and subsets created by the users. Private views and subsets are kept in folders with the user name of each individual user. When migrating to CloudPak, the user name will change to using IBM IDs and the following steps will migrate the user’s private views and subsets using their new IBM IDs.
- In the }ClientProperties control dimension, add an element nam ed CloudID.
- In the }ClientProperties control cube, add its new Planning Analytics Engine name to the CloudID cell for each client that you wish to migrate. Private views and subsets will be migrated under the specified CloudID names. Clients without a defined CloudID names will not be migrated.
Migrating Database by Creating tar.gz File
You need to download the latest IBM TM1 to Planning Analytics Engine Database Migration Utility tool from the IBM Fix Central, from Passport Advantage or from your new PAW environment. Download Git Bash Windows (if you are currently running your Planning Analytics on Windows).
- Install the Git Bash application on the server where the current Planning Analytics Local is hosted. Run the appropriate command to create a backup of your database.
| Platform | Command |
| Linux | migrate_tm1.sh –input /models/X –output file://localhost/xfer/PlanSamp.tar.gz |
| Windows | ./migrate_tm1.sh –input “C:\ibm\models\PlanSamp” –output “file://localhost/c:/xfer/PlanSamp.tar.gz” |
- Copy the migration file to the Planning Analytics Engine, then upload in Databases under Administration.
- Make sure that the server or database in Planning Analytics Engine is not running. Click on the 3 dots on the right side of the pane and click Restore.
Need Help Migrating to Planning Analytics as a Service?
If you have any questions about migrating from On-Prem TM1 to Planning Analytics as a Service, please reach out to us. We’d be delighted to help.
