On 13/04/17 05:04, Euler Taveira wrote:
> Hi,
>
> If a certain table has different schemas and the subscriber table has an
> unmatched column with a not null constraint, the logical replication
> crashes with the above stack trace.
>
> [snip]
>
> Are we prepared to support different schemas in v10? Or should we
> disallow it for v10 and add a TODO?
>
Ah nuts, yes it's supposed to be supported, we seem to not initialize
cstate->range_table in tablesync which causes this bug. The CopyState
struct is private to copy.c so we can't easily set cstate->range_table
externally. I wonder if tablesync should just construct CopyStmt instead
of calling the lower level API.
-- Petr Jelinek http://www.2ndQuadrant.com/ PostgreSQL Development, 24x7 Support, Training &
Services