Bug#26464 - insert delayed + update + merge = corruption

Using INSERT DELAYED on MERGE tables could lead to table
corruptions.

The manual lists a couple of storage engines, which can be
used with INSERT DELAYED. MERGE is not in this list.

The attempt to try it anyway has not been rejected yet.
This bug was not detected earlier as it can work under
special circumstances. Most notable is low concurrency.

To be safe, this patch rejects any attempt to use INSERT
DELAYED on MERGE tables.
parent cf9aca84
......@@ -801,3 +801,8 @@ CREATE TABLE tm1(a SMALLINT, b SMALLINT, KEY(a)) ENGINE=MERGE UNION=(t1);
SELECT * FROM tm1;
ERROR HY000: Unable to open underlying table which is differently defined or of non-MyISAM type or doesn't exist
DROP TABLE t1, tm1;
CREATE TABLE t1(c1 INT) ENGINE=MyISAM;
CREATE TABLE t2(c1 INT) ENGINE=MERGE UNION=(t1);
INSERT DELAYED INTO t2 VALUES(1);
ERROR HY000: Table storage engine for 't2' doesn't have this option
DROP TABLE t1, t2;
......@@ -428,4 +428,13 @@ CREATE TABLE tm1(a SMALLINT, b SMALLINT, KEY(a)) ENGINE=MERGE UNION=(t1);
SELECT * FROM tm1;
DROP TABLE t1, tm1;
#
# Bug#26464 - insert delayed + update + merge = corruption
#
CREATE TABLE t1(c1 INT) ENGINE=MyISAM;
CREATE TABLE t2(c1 INT) ENGINE=MERGE UNION=(t1);
--error 1031
INSERT DELAYED INTO t2 VALUES(1);
DROP TABLE t1, t2;
# End of 4.1 tests
......@@ -37,7 +37,7 @@ class ha_myisammrg: public handler
{
return (HA_REC_NOT_IN_SEQ | HA_AUTO_PART_KEY | HA_READ_RND_SAME |
HA_NULL_IN_KEY | HA_CAN_INDEX_BLOBS | HA_FILE_BASED |
HA_CAN_INSERT_DELAYED | HA_ANY_INDEX_MAY_BE_UNIQUE);
HA_ANY_INDEX_MAY_BE_UNIQUE);
}
ulong index_flags(uint inx, uint part, bool all_parts) const
{
......
Markdown is supported
0%
or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment