Bug #48883 Test "innodb_information_schema" takes fewer locks than expected
Submitted: 18 Nov 2009 21:13 Modified: 17 Aug 2011 14:48
Reporter: Joerg Bruehe Email Updates:
Status: Closed Impact on me:
None 
Category:MySQL Server: Tests Severity:S3 (Non-critical)
Version:5.1, 5.5, 5.6 OS:Linux
Assigned to: Vasil Dimov
Tags: experimental, pb2, test failure
Triage: Triaged: D3 (Medium) / R1 (None/Negligible) / E1 (None/Negligible)

[18 Nov 2009 21:13] Joerg Bruehe
Description:
Observer during the build of 5.5.0-beta.

It has not been observed before, but that may well be due to the recent change in the test. From the effect I see (and from bug DB entries), this source contains the fix for bug#47808.

I'm not sure whether this is a bug in the server, specific in InnoDB, or in the test.

Symptom:
=====
innodb.innodb_information_schema         [ retry-fail ]
        Test ended at YYYY-MM-DD HH:MM:SS

CURRENT_TEST: innodb.innodb_information_schema
--- /PATH/mysql-test/suite/innodb/r/innodb_information_schema.result
+++ /PATH/mysql-test/suite/innodb/r/innodb_information_schema.reject
@@ -1,8 +1,7 @@
+Timeout in wait_condition.inc for select count(*) = 14 from INFORMATION_SCHEMA.INNODB_LOCKS
 lock_mode      lock_type       lock_table      lock_index      lock_rec        lock_data
 X      RECORD  `test`.```t'\"_str`     `PRIMARY`       2       '1', 'abc', '''abc', 'abc''', 'a''bc', 'a''bc'
'', '''abc'''''
 X      RECORD  `test`.```t'\"_str`     `PRIMARY`       2       '1', 'abc', '''abc', 'abc''', 'a''bc', 'a''bc'
'', '''abc'''''
-X      RECORD  `test`.```t'\"_str`     `PRIMARY`       3       '2', 'abc', '"abc', 'abc"', 'a"bc', 'a"bc"', '
"abc""'
-X      RECORD  `test`.```t'\"_str`     `PRIMARY`       3       '2', 'abc', '"abc', 'abc"', 'a"bc', 'a"bc"', '
"abc""'
 X      RECORD  `test`.```t'\"_str`     `PRIMARY`       4       '3', 'abc', '\\abc', 'abc\\', 'a\\bc', 'a\\bc\
\', '\\abc\\\\'
 X      RECORD  `test`.```t'\"_str`     `PRIMARY`       4       '3', 'abc', '\\abc', 'abc\\', 'a\\bc', 'a\\bc\
\', '\\abc\\\\'
 X      RECORD  `test`.```t'\"_str`     `PRIMARY`       5       '4', 'abc', '\0abc', 'abc\0', 'a\0bc', 'a\0bc\
0', 'a\0bc\0\0'
@@ -16,8 +15,8 @@
 lock_table     COUNT(*)
 `test`.`t_max` 2
 `test`.`t_min` 2
-`test`.```t'\"_str`    10
+`test`.```t'\"_str`    8
 lock_table     COUNT(*)
 "test"."t_max" 2
 "test"."t_min" 2
-"test"."`t'\""_str"    10
+"test"."`t'\""_str"    8

mysqltest: Result content mismatch
=====
RPM build for RHEL 3, x86, normal test mode.

May be stronger ("timeout" lines not pasted):
=====
...
@@ -16,8 +13,8 @@
 lock_table     COUNT(*)
 `test`.`t_max` 2
 `test`.`t_min` 2
-`test`.```t'\"_str`    10
+`test`.```t'\"_str`    6
 lock_table     COUNT(*)
 "test"."t_max" 2
 "test"."t_min" 2
-"test"."`t'\""_str"    10
+"test"."`t'\""_str"    6
=====
Generic RPM for x86, test run in PS mode.

Or even more (again, "timeout" lines not pasted):
=====
...
@@ -16,8 +11,8 @@
 lock_table     COUNT(*)
 `test`.`t_max` 2
 `test`.`t_min` 2
-`test`.```t'\"_str`    10
+`test`.```t'\"_str`    4
 lock_table     COUNT(*)
 "test"."t_max" 2
 "test"."t_min" 2
-"test"."`t'\""_str"    10
+"test"."`t'\""_str"    4
=====
RPM build for RHEL 3, x86_64, debug tests,  and also
Generic RPM for x86, test run in normal mode.

I have no idea why it was observed only in RPM builds,
I am not aware of any significant difference with the machines or the way tests are run compared to a tar.gz build.

How to repeat:
Found while running the test suite.
[28 Dec 2009 0:00] Bugs System
No feedback was provided for this bug for over a month, so it is
being suspended automatically. If you are able to provide the
information that was originally requested, please do so and change
the status of the bug back to "Open".
[11 May 2010 12:03] Alexander Nozdrin
The test case has been marked experimental starting from mysql-trunk due to this bug.
[17 Feb 2011 17:22] Vasil Dimov
Test non-determinism fixed in vasil.dimov@oracle.com-20110217170953-952xbu45ww1rpl46