Heap Allocation Error - Large Varchar Field

[Raftsman]

We have an input staging table that contains a column with a Varchar2(4000). We have had an increase in rows from 10000 to ~60,000. We are running under 1 node.

Question: When Datastage reads the input, I assume it reads the column as 4000 even though it's variable. Am I correct?

I increased to 2 nodes and my problem went away. From a scalability standpoint is there a best practice to follow when dealing with large data type lengths.

Thanks