[MDEV-19454] rpl.rpl_truncate_3innodb failed in buildbot with syntax error in test Created: 2019-05-13  Updated: 2023-11-29  Resolved: 2023-11-29

Status: Closed
Project: MariaDB Server
Component/s: Tests
Affects Version/s: 10.1
Fix Version/s: N/A

Type: Bug Priority: Major
Reporter: Elena Stepanova Assignee: Angelique Sklavounos (Inactive)
Resolution: Won't Fix Votes: 0
Labels: None


 Description   

http://buildbot.askmonty.org/buildbot/builders/kvm-deb-bionic-aarch64/builds/1889/steps/mtr/logs/stdio

10.1 d0ee3b55

rpl.rpl_truncate_3innodb 'mix,xtradb'    w4 [ fail ]
        Test ended at 2019-05-09 13:17:08
 
CURRENT_TEST: rpl.rpl_truncate_3innodb
mysqltest: In included file "./include/show_rpl_debug_info.inc": 
included from ./include/wait_for_slave_param.inc at line 115:
included from ./include/wait_for_slave_io_to_start.inc at line 40:
included from ./include/wait_for_slave_to_start.inc at line 26:
included from ./include/start_slave.inc at line 35:
included from ./include/rpl_for_each_slave.inc at line 33:
included from ./include/rpl_start_slaves.inc at line 30:
included from ./include/rpl_init.inc at line 210:
included from ./include/master-slave.inc at line 38:
included from ./extra/rpl_tests/rpl_truncate.test at line 6:
included from /usr/share/mysql/mysql-test/suite/rpl/t/rpl_truncate_3innodb.test at line 4:
At line 93: Found line beginning with --  that didn't contain a valid mysqltest command, check your syntax or use # if you intended to write a comment
 
The result from queries just before the failure was:
include/master-slave.inc
 
 - saving '/dev/shm/var/4/log/rpl.rpl_truncate_3innodb-mix,xtradb/' to '/dev/shm/var/log/rpl.rpl_truncate_3innodb-mix,xtradb/'
 
Retrying test rpl.rpl_truncate_3innodb, attempt(2/3)...
 
worker[4] > Restart  - not started
worker[4] > Restart  - not started
rpl.rpl_truncate_3innodb 'mix,xtradb'    w4 [ retry-pass ]    964



 Comments   
Comment by Angelique Sklavounos (Inactive) [ 2022-03-02 ]

Having looked at the code, there is seemingly not an obvious issue (to me) with the syntax:
https://github.com/MariaDB/server/blob/d0ee3b5500cdaf3382fc8f26c883f45d8f4bdf5b/mysql-test/include/show_rpl_debug_info.inc#L93

This same error was also seen in MDEV-25826:
https://github.com/MariaDB/server/blob/5bd517259f2bb3e48cd3dab/mysql-test/include/end_include_file.inc#L55

A search on Multi-CI cross-reference shows this was only occurrence of this type of fail with this test. However, a search more generally for “Found line beginning with” in Failure Output produced similar fails:

10.3 25d6bbcd5 kvm-deb-focal-aarch64 rpl.rpl_auto_increment innodb,stmt nm Normal run, no --ps-protocol 2021-09-21 20:49:45
https://github.com/MariaDB/server/blob/25d6bbcd5172eb9af0a68be11fb66245/mysql-test/include/show_rpl_debug_info.inc#L59

rpl.rpl_auto_increment 'innodb,stmt'     w2 [ fail ]
        Test ended at 2021-09-21 21:00:13
 
CURRENT_TEST: rpl.rpl_auto_increment
mysqltest: In included file "./include/show_rpl_debug_info.inc": 
included from ./include/wait_for_slave_param.inc at line 81:
included from ./include/wait_for_slave_io_to_stop.inc at line 36:
included from ./include/wait_for_slave_to_stop.inc at line 35:
included from ./include/stop_slave.inc at line 84:
included from ./include/rpl_for_each_slave.inc at line 33:
included from ./include/rpl_stop_slaves.inc at line 31:
included from ./include/rpl_end.inc at line 81:
included from /usr/share/mysql/mysql-test/suite/rpl/include/rpl_auto_increment.test at line 318:
included from /usr/share/mysql/mysql-test/suite/rpl/t/rpl_auto_increment.test at line 8:
At line 59: Found line beginning with --  that didn't contain a valid mysqltest command, check your syntax or use # if you intended to write a comment
 
The result from queries just before the failure was:
< snip >
BEGIN;
# Set sql_mode with NO_AUTO_VALUE_ON_ZERO for allowing
# zero to fill the auto_increment field.
SET SQL_MODE=NO_AUTO_VALUE_ON_ZERO;
INSERT INTO t1(id,data) VALUES(0,2);
# Resetting sql_mode without NO_AUTO_VALUE_ON_ZERO to
# affect the execution of the transaction on slave.
SET SQL_MODE=0;
COMMIT;
SELECT * FROM t1;
id	data
0	2
connection slave;
SELECT * FROM t1;
id	data
0	2
connection master;
DROP TABLE t1;
connection slave;
include/rpl_end.inc

10.2 dd5c307cb kvm-deb-bionic-aarch64 rpl.rpl_row_001 row nm Normal run, no --ps-protocol 2020-04-29 08:52:27
https://github.com/MariaDB/server/blob/dd5c307cb00bfde6c88bf125f61b3e0d85dc79a5/mysql-test/include/end_include_file.inc#L71

rpl.rpl_row_001 'row'                    w4 [ fail ]
        Test ended at 2020-04-29 05:07:20
 
CURRENT_TEST: rpl.rpl_row_001
mysqltest: In included file "./include/end_include_file.inc": 
included from ./include/check_slave_no_error.inc at line 32:
included from ./include/rpl_end.inc at line 74:
included from /usr/share/mysql/mysql-test/suite/rpl/t/rpl_row_001.test at line 12:
At line 71: Found line beginning with --  that didn't contain a valid mysqltest command, check your syntax or use # if you intended to write a comment
 
The result from queries just before the failure was:
< snip >
5000
connection slave;
LOCK TABLES t1 READ;
START SLAVE;
UNLOCK TABLES;
SELECT COUNT(*) FROM t1;
COUNT(*)
5000
connection master;
DROP TABLE t1;
CREATE TABLE t1 (n INT);
INSERT INTO t1 VALUES(3456);
connection slave;
SELECT n FROM t1;
n
3456
connection master;
DROP TABLE t1;
connection slave;
include/rpl_end.inc

In the above, show_rpl_debug_info.inc was called by wait_for_slave_param.inc at lines 115 and 81 (or rather the lines above), where either a timeout or misconfiguration occurs:
https://github.com/MariaDB/server/blob/d0ee3b5500cdaf3382fc8f26c883f45d8f4bdf5b/mysql-test/include/wait_for_slave_param.inc#L115
https://github.com/MariaDB/server/blob/25d6bbcd5172eb9af0a68be11fb66245/mysql-test/include/wait_for_slave_param.inc#L81

And end_include_file.inc was called by wait_for_slave_sql_to_start.inc at line 38 and check_slave_no_error at line 32:
https://github.com/MariaDB/server/blob/5bd517259f2bb3e48cd3dab/mysql-test/include/wait_for_slave_sql_to_start.inc#L37
https://github.com/MariaDB/server/blob/dd5c307cb00bfde6c88bf125f61b3e0d85dc79a5/mysql-test/include/check_slave_no_error.inc#L31

The failure hasn’t happened very often, but it seems like something to investigate further by trying to manually trigger the error.

Generated at Thu Feb 08 08:51:49 UTC 2024 using Jira 8.20.16#820016-sha1:9d11dbea5f4be3d4cc21f03a88dd11d8c8687422.