Details
-
Bug
-
Status: Closed (View Workflow)
-
Critical
-
Resolution: Fixed
-
10.1(EOL), 10.2(EOL), 10.3(EOL), 10.4(EOL), 10.5, 10.6, 10.7(EOL), 10.8(EOL), 10.9(EOL), 10.10(EOL), 10.11, 11.0(EOL)
Description
Most probably this is this relatively new test, introduced by dbe447a7890 internal issue.
To be fixed soon.
Attachments
Issue Links
- relates to
-
-
- Closed
-
Activity
https://buildbot.mariadb.org/#/builders/192/builds/7772/steps/6/logs/stdio
10.8 1c5b099a9619c953e7510bbafca89353ad0a020c
https://buildbot.mariadb.org/#/builders/192/builds/7783/steps/6/logs/stdio
10.9 bbe99cd4e2d7c83a06dd93ea88af97f2d5796810
I ran into this also. Sporadic, but common, failure. Issue occurrence may be more pronounced after bb-10.4-MDEV-30619.
|
bb-10.4-MDEV-30619 10.4.31 006aca3111e0cbd5f172060774c74e328bea4b0a (Debug) |
...
|
rpl.rpl_parallel_optimistic_until 'innodb,row' w21 [ 157 fail ]
|
Test ended at 2023-07-22 17:25:26
|
|
|
CURRENT_TEST: rpl.rpl_parallel_optimistic_until
|
mysqltest: At line 274: Query 'SHOW RELAYLOG EVENTS IN 'slave-relay-bin.000003' FROM 508 LIMIT 1' didn't return a result set
|
|
|
The result from queries just before the failure was:
|
< snip >
|
1
|
SELECT count(*) = 0 as 'trx2 is not committed' FROM t1 WHERE b = 'trx2';
|
trx2 is not committed
|
1
|
Proof 2: Resume works out
|
include/start_slave.inc
|
connection master;
|
connection slave;
|
*** case 4 Relay-log UNTIL inside trx1
|
connection slave;
|
DELETE FROM t1 WHERE a <> 25;
|
UPDATE t1 SET b='trx0' WHERE a = 25;
|
connection slave1;
|
BEGIN;
|
INSERT INTO t1 SET a= 1; # block trx1;
|
connection slave;
|
include/stop_slave.inc
|
CHANGE MASTER TO MASTER_LOG_FILE = 'file_1', MASTER_LOG_POS = <pos_trx0>, MASTER_USE_GTID=no;
|
START SLAVE IO_THREAD;
|
include/wait_for_slave_io_to_start.inc
|
Also observed on same branch, also sporadic, though (much) less frequent:
|
10.4.31 006aca3111e0cbd5f172060774c74e328bea4b0a (Debug) |
rpl.rpl_parallel_optimistic_until 'innodb,row' w43 [ 207 fail ]
|
Test ended at 2023-07-23 16:56:06
|
|
|
CURRENT_TEST: rpl.rpl_parallel_optimistic_until
|
mysqltest: At line 279:
|
|
|
The result from queries just before the failure was:
|
< snip >
|
SELECT count(*) = 0 as 'trx2 is not committed' FROM t1 WHERE b = 'trx2';
|
trx2 is not committed
|
1
|
Proof 2: Resume works out
|
include/start_slave.inc
|
connection master;
|
connection slave;
|
*** case 4 Relay-log UNTIL inside trx1
|
connection slave;
|
DELETE FROM t1 WHERE a <> 25;
|
UPDATE t1 SET b='trx0' WHERE a = 25;
|
connection slave1;
|
BEGIN;
|
INSERT INTO t1 SET a= 1; # block trx1;
|
connection slave;
|
include/stop_slave.inc
|
CHANGE MASTER TO MASTER_LOG_FILE = 'file_1', MASTER_LOG_POS = <pos_trx0>, MASTER_USE_GTID=no;
|
START SLAVE IO_THREAD;
|
include/wait_for_slave_io_to_start.inc
|
*** Unexpected offset. Refine it to point to the correct XID event!
|
10.7 d7af7bfc2b6ab89801059406e1bd74846a2ca8e8
rpl.rpl_parallel_optimistic_until 'innodb,row' w8 [ fail ]Test ended at 2021-10-28 09:27:10CURRENT_TEST: rpl.rpl_parallel_optimistic_untilmysqltest: At line 274: Query 'SHOW RELAYLOG EVENTS IN 'slave-relay-bin.000003' FROM 508 LIMIT 1' didn't return a result set