Ability to add columns to DB2 z/OS tables without impacting IDAA or incremental update
Currently, adding a column to a DB2 z/OS table that is using IDAA with incremental update involves:
1) Adding the column in DB2 z/OS
2) Stopping replication for all IDAA tables
3) Dropping the affected table from IDAA
4) Adding the affected table back to IDAA
5) Restarting replication, which first forces a full reload of the affected table.
I am requesting that this process be simplified to:
1) Add the new column to the DB2 z/OS table
That's it. There should be no need to drop/add tables to IDAA. There should be no need to manually stop/start replication, and there should be no need to do a reload.
With CDC from DB2 z/OS to other types of target databases, we accomplish this with the following steps. IDAA should automate most of this. It does require ADDCOLUMNISSCHEMACHANGE=YES and ONSCHEMACHANGE=STOP parameters for CDC z/OS.
1 - Applications touching the table to modified are quiesced. (Manual process)
2- The add column is performed to the DB2 z/OS table. (Manual process)
3- When CDC sees the add column in the DB log, it stops replication for the CDC subscription. (Automatic within CDC today)
5 - The target table is altered to add the same column (IDAA should automate this)
6 - A refresh of the CDC metadata for the source and target tables is performed. (this function is available in the CDC access server java API and this should be automated by IDAA)
7 - The subscription is restarted. (Should be automated by IDAA).
8 - Applications are allowed back in. (Manual process - but I would suggest that the subscription start message could be trapped by the customer to know when everything is ready for application access)