Ubuntu GNU/Linux 18.04 (Bionic Beaver) x64

It has been noted that INSERT can chuck errors for reasons other than a duplicate key.
As the manual notes
"IGNORE has a similar effect on inserts into partitioned tables where no partition matching a given value is found. Without IGNORE, such INSERT statements are aborted with an error. When INSERT IGNORE is used, the insert operation fails silently for rows containing the unmatched value, but inserts rows that are matched."

One example would be a foreign key check failing, in this case IGNORE would see the INSERT return without error (but without inserting a row, of course).

However, when combined with ON DUPLICATE KEY may create an ambiguity.
if an fk check fails, but there is also a dup key, which one takes precedence?
That is, will the on dup key be executed in that case if IGNORE is also specified, or will the server just return without error because of the fk check fail?
If we intend to allow this syntax, then the exact behaviour for each of the possible combinations of errors should be documented.

IGNORE can also play a role even in a pure duplicate key situation. Consider the following:

MariaDB [test]> create table t1 (pk int primary key, c varchar(8),unique(c));

Query OK, 0 rows affected (0.04 sec)

MariaDB [test]> insert into t1 values (1,'foo'),(2,'bar');

Query OK, 2 rows affected (0.00 sec)

Records: 2  Duplicates: 0  Warnings: 0

MariaDB [test]> insert into t1 values (1,'qux') on duplicate key update c = 'bar';

ERROR 1062 (23000): Duplicate entry 'bar' for key 'c'

MariaDB [test]> insert ignore into t1 values (1,'qux') on duplicate key update c = 'bar';

Query OK, 0 rows affected, 1 warning (0.00 sec)

MariaDB [test]> show warnings;

+---------+------+-----------------------------------+

| Level   | Code | Message                           |

+---------+------+-----------------------------------+

| Warning | 1062 | Duplicate entry 'bar' for key 'c' |

+---------+------+-----------------------------------+

1 row in set (0.00 sec)

MariaDB [test]> select * from t1;

+----+------+

| pk | c    |

+----+------+

|  1 | foo  |

|  2 | bar  |

+----+------+

2 rows in set (0.00 sec)

I agree the behaviour should be properly documented.

This is documented as follows "The IGNORE and DELAYED options are ignored when you use ON DUPLICATE KEY UPDATE." at https://mariadb.com/kb/en/library/insert-on-duplicate-key-update/
I have added to https://mariadb.com/kb/en/library/insert-ignore/ as well.