Bug #26377 Deadlock with MERGE and FLUSH TABLE
Submitted: 14 Feb 2007 18:46 Modified: 3 Dec 2007 9:33
Reporter: Dean Ellis Email Updates:
Status: Closed Impact on me:
None 
Category:MySQL Server: Merge storage engine Severity:S1 (Critical)
Version:4.0 4.1 5.0 5.1 OS:Any
Assigned to: Ingo Strüwing CPU Architecture:Any
Tags: bfsm_2007_03_01

[14 Feb 2007 18:46] Dean Ellis
Description:
One can deadlock the server with a FLUSH TABLE statement following a LOCK TABLES statement with the tables (involving a MERGE table) listed in a particular order.

How to repeat:
This does not deadlock:

DROP TABLE IF EXISTS m1, t1;
CREATE TABLE t1 ( a INT ) ENGINE=MyISAM;
CREATE TABLE m1 ( a INT ) ENGINE=MERGE UNION=(t1);
LOCK TABLES t1 WRITE, m1 WRITE;
FLUSH TABLE t1;
UNLOCK TABLES;

This does deadlock:

DROP TABLE IF EXISTS m1, t1;
CREATE TABLE t1 ( a INT ) ENGINE=MyISAM;
CREATE TABLE m1 ( a INT ) ENGINE=MERGE UNION=(t1);
LOCK TABLES m1 WRITE, t1 WRITE;
FLUSH TABLE t1;
UNLOCK TABLES;

Suggested fix:
Heh.
[16 Apr 2007 16:23] Sergey Vojtovich
Back to verified as no progress on this bug so far.
[2 May 2007 10:38] Ingo Strüwing
This is fixed by my proposed fix for Bug#26379. If it becomes approved, I'll close this bug too.
[16 Nov 2007 8:08] Ingo Strüwing
For the patch see Bug#26379 - Combination of FLUSH TABLE and REPAIR TABLE corrupts a MERGE table.
[28 Nov 2007 10:34] Ingo Strüwing
For documentation see Bug#26379 - Combination of FLUSH TABLE and REPAIR TABLE corrupts a MERGE table.
[3 Dec 2007 9:33] MC Brown
A note has been added to the 5.1.23 and 6.0.4 changelogs: 

Important Change: Incompatible Change: A number of problems existed in the implementation of MERGE tables that could cause problems. The problems are summarized below:

Bug#26379 - Combination of FLUSH TABLE and REPAIR TABLE corrupts a MERGE table. This was caused in a number of situations:

A thread trying to lock a MERGE table performs busy waiting while REPAIR TABLE or a similar table administration task is ongoing on one or more of its MyISAM tables.

A thread trying to lock a MERGE table performs busy waiting until all threads that did REPAIR TABLE or similar table administration tasks on one or more of its MyISAM tables in LOCK TABLES segments do UNLOCK TABLES. The difference against problem #1 is that the busy waiting takes place after the administration task. It is terminated by UNLOCK TABLES only.

Two FLUSH TABLES within a LOCK TABLES segment can invalidate the lock. This does not require a MERGE table. The first FLUSH TABLES can be replaced by any statement that requires other threads to reopen the table. In 5.0 and 5.1 a single FLUSH TABLES can provoke the problem.

Bug#26867 - Simultaneously executing LOCK TABLES and REPAIR TABLE on a MERGE table would result in memory/cpu hogging.

Trying DML on a MERGE table, which has a child locked and repaired by another thread, made an infinite loop in the server.

Bug#26377 - Deadlock with MERGE and FLUSH TABLE

Locking a MERGE table and its children in parent-child order and flushing the child deadlocked the server.

Bug#25038 - Waiting TRUNCATE

Truncating a MERGE child, while the MERGE table was in use, let the truncate fail instead of waiting for the table to become free.

Bug#25700 - MERGE base tables get corrupted by OPTIMIZE/ANALYZE/REPAIR TABLE

Repairing a child of an open MERGE table corrupted the child. It was necessary to FLUSH the child first.

Bug#30275 - MERGE tables: FLUSH TABLES or UNLOCK TABLES causes server to crash.

Flushing and optimizing locked MERGE children crashed the server.

Bug#19627 - temporary merge table locking

Use of a temporary MERGE table with non-temporary children could corrupt the children.

Temporary tables are never locked. Creation of tables with non-temporary chidlren of a temporary MERGE table is now prohibited.

Bug#27660 - Falcon: MERGE table possible

It was possible to create a MERGE table with non-MyISAM children.

Bug#30273 - MERGE tables: Can't lock file (errno: 155)

This was a Windows-only bug. Table administration statements sometimes failed with "Can't lock file (errno: 155)".

The fix introduces the following changes in behavior:

This patch changes the behavior of temporary MERGE tables. Temporary MERGE must have temporary children. The old behavior was wrong. A temporary table is not locked. Hence even non-temporary children were not locked. See Bug#19627.

You cannot change the union list of a non-temporary MERGE table when LOCK TABLES is in effect. The following does not work:

CREATE TABLE m1 ... ENGINE=MRG_MYISAM ...;
      LOCK TABLES t1 WRITE, t2 WRITE, m1 WRITE;
      ALTER TABLE m1 ... UNION=(t1,t2) ...;
However, you can do this with a temporary MERGE table.

You cannot create a MERGE table with CREATE ... SELECT, neither as a temporary MERGE table, nor as a non-temporary MERGE table. For example:

CREATE TABLE m1 ... ENGINE=MRG_MYISAM ... SELECT ...;
Gives error message: table is not BASE TABLE.