How to specify multiple Goldengate GGSCHEMA settings for multi-tenant deployments?

Given the following:

Command Line Interface Reference for Oracle GoldenGate https://docs.oracle.com/en/middleware/goldengate/core/21.1/gclir/add-heartbeattable.html#GUID-126E30A2-DC7A-4C93-93EC-0EB8BA7C13CB Use ADD HEARTBEATTABLE to create the objects necessary to use the automatic heartbeat functionality. This command performs the following tasks
<https://docs.oracle.com/en/middleware/goldengate/core/21.1/gclir/add-heartbeattable.html#GUID-126E30A2-DC7A-4C93-93EC-0EB8BA7C13CB>

• Extract: Use the schema name in each PDB. Each PDB with which the Extract is registered should have its own heartbeat table. To instruct Extract which schema to find the heartbeat tables in, use GGSCHEMA pdb.schemaname within the GLOBALS file.

and....

GGSCHEMA
https://docs.oracle.com/goldengate/c1230/gg-winux/GWURF/ggschema.htm#GWURF517 Use the GGSCHEMA parameter to specify the name of the schema that contains the database objects that are owned by Oracle GoldenGate, such as those that support DDL replication for trigger-based replication, those that are a part of the heartbeat table implementation, and those that are part of the SQL Server CDC Capture and Cleanup implementation..
<https://docs.oracle.com/goldengate/c1230/gg-winux/GWURF/ggschema.htm#GWURF517>

*Syntax*

GGSCHEMA [container.]schema_name

Given the above how exactly do you go about specifying multiple GGSCHEMAS for use with multitenant? What if I want to use goldengate across a mixture of non-CDB and CDB sources?

I tried both with multiple GGSCHEMA parameters and comma-separated schemas, getting errors both times.

-- 
Charles Schultz

--
http://www.freelists.org/webpage/oracle-l