After adding a column family you must code your app to wait until schema versions became 1 in the cluster to add a new row. This is something that must be solved in the application layer.
In pycassa (python api) the system manager code already do this. I don't know about others clients.
I add a column family dynamically and notice that when describe schema versions return 2 values. Then it quickly changes back to 1. Sometimes this stays at 2 and does not change. Then I cannot insert values to the created column family, as it causes an exception.
Is there a way to force the schema propagation through the thrift API (not the CLI)?
Thanks in advance.