Re: INSERT INTO schema.table VS use schema

[Rick James]

Well, it is either a bug or a feature. It does case grief. Here's the story:

If

1. You have some kind of filtering: binlog/replication_do/ignore_... in the my.cnf on the Master and/or Slave. And...

2. You have the table name qualified by the database (test.animals) in the write query (INSERT). And...

3. The default database is not the same as the qualification (`test` in your example). (For example, you had "USE foo".) And...

4. The default db (foo) is not replicated, but the qualification (test) is. Then...

the write will be mysteriously not replicated.

Unravel that to see what you can do to avoid the "problem"...
* Don't use binlog/replication_do/ignore.
* Don't use db qualifiers.
* Be sure to USE the right db (as you saw with the case that worked).

Since you are coming through a third-party API (Drupal), you may not have control over those workarounds. Complain to them.