Details
-
Bug
-
Status: Stalled (View Workflow)
-
Critical
-
Resolution: Unresolved
-
11.7(EOL)
-
None
Description
The test case is very non-deterministic, but it's remarkably well rr-able (it can take a few thousands attempts on my machine without rr, but usually fails in a few attempts with --rr=-h).
Try to run with --repeat=N and --rr=-h.
It looks like the failure started happening on 11.7 after the merge:
commit b12ff287ecefef4213e4e24c4413e7c63ad1f84a 7aa28a2a546620bbcc066704f0e6cf2fc007de6d d8dad8c3b54cd09fefce7bc3b9749f427eed9709
|
Merge: 7aa28a2a546 d8dad8c3b54
|
Commit: Oleksandr Byelkin <sanja@mariadb.com>
|
CommitDate: Sun Nov 10 19:22:21 2024 +0100
|
|
|
Merge branch '11.6' into 11.7
|
However, I couldn't reproduce it on 11.6. Also, since the test case behaves generally differently without innnodb_snapshot_isolation, the commit enabling innnodb_snapshot_isolation by default seems a likely suspect, but the failure isn't reproducible on 11.7 before the merge above even with the explicit innnodb_snapshot_isolation=on. There weren't many other 11.6 changes in that merge, so possibly the difference comes from lower versions.
--source include/have_innodb.inc
|
--source include/have_sequence.inc
|
|
|
SET @isolation.save= @@global.innodb_snapshot_isolation; |
SET GLOBAL innodb_snapshot_isolation= ON; |
|
|
CREATE TABLE t1 (a INT) ENGINE=InnoDB; |
CREATE TABLE t2 (col_char CHAR(224), id SERIAL) ENGINE=InnoDB; |
CREATE TABLE t3 ( |
id int(10) unsigned NOT NULL auto_increment, |
c char(120) NOT NULL default '', |
PRIMARY KEY (id)) ENGINE=Aria; |
INSERT INTO t3 SELECT seq, CHR(seq%26+97) FROM seq_1_to_100; |
|
|
--connect (con1,localhost,root,,)
|
--connect (con2,localhost,root,,)
|
--connect (con3,localhost,root,,)
|
XA BEGIN 'x1'; |
INSERT INTO t2 () VALUES (); |
|
|
--connection con2
|
XA START 'x2'; |
SELECT * FROM t1; |
--error ER_CHECKREAD
|
UPDATE t2 SET col_char = NULL; |
|
|
--connection con1
|
--send
|
ALTER TABLE t3 ADD KEY (c); |
|
|
--connection con2
|
DELETE FROM t3 WHERE id = 0; |
UPDATE t3 SET c = ''; |
--error ER_XA_RBROLLBACK
|
XA END 'x2'; |
XA ROLLBACK 'x2'; |
--connection con1
|
--reap
|
--disconnect con1
|
--disconnect con2
|
--connection con3
|
XA END 'x1'; |
XA ROLLBACK 'x1'; |
--disconnect con3
|
|
|
--connection default
|
DROP TABLE t1, t2, t3; |
SET GLOBAL innodb_snapshot_isolation= @isolation.save; |
2024-12-16 15:54:40 6 [Warning] WSREP: handlerton rollback failed, thd 6 48 conf 0 SQL XA ROLLBACK 'x2'
|
mariadbd: /data/MDEV-35661/11.7/sql/online_alter.cc:426: int online_alter_close_connection(THD*): Assertion `!cache_list || cache_list->empty()' failed.
|
|
|
#3 0x000048f350d11fd6 in __GI___assert_fail (assertion=0x55f6effd0a80 "!cache_list || cache_list->empty()",
|
file=0x55f6effd0810 "/data/MDEV-35661/11.7/sql/online_alter.cc", line=426, function=0x55f6effd0aa8 "int online_alter_close_connection(THD*)")
|
at assert.c:101
|
#4 0x000055f6ef4b9656 in online_alter_close_connection (thd=0x5abcfc000d48) at /data/MDEV-35661/11.7/sql/online_alter.cc:426
|
#5 0x000055f6ef2a8023 in ha_close_connection (thd=0x5abcfc000d48) at /data/MDEV-35661/11.7/sql/handler.cc:967
|
#6 0x000055f6eee3ad88 in THD::free_connection (this=0x5abcfc000d48) at /data/MDEV-35661/11.7/sql/sql_class.cc:1732
|
#7 0x000055f6eed3c32e in unlink_thd (thd=0x5abcfc000d48) at /data/MDEV-35661/11.7/sql/mysqld.cc:2876
|
#8 0x000055f6ef0944dc in do_handle_one_connection (connect=0x55f6f231e6f8, put_in_cache=true) at /data/MDEV-35661/11.7/sql/sql_connect.cc:1426
|
#9 0x000055f6ef094211 in handle_one_connection (arg=0x55f6f231e6f8) at /data/MDEV-35661/11.7/sql/sql_connect.cc:1327
|
#10 0x00005abd001f8609 in start_thread (arg=<optimized out>) at pthread_create.c:477
|
#11 0x000048f350dfd353 in clone () at ../sysdeps/unix/sysv/linux/x86_64/clone.S:95
|