lp:~tsarev/percona-server/5.5_slow_extended_tests_fixes

Created by Oleg Tsarev on 2011-07-11 and last modified on 2011-07-13
Get this branch:
bzr branch lp:~tsarev/percona-server/5.5_slow_extended_tests_fixes
Only Oleg Tsarev can upload to this branch. If you are Oleg Tsarev please log in for upload directions.

Branch merges

Related bugs

Related blueprints

Branch information

Recent revisions

138. By Oleg Tsarev on 2011-07-13

1) Move all tests from mysql-test/slow_extended.patch to slow_extended.patch.
Now tests in "percona" suite
2) Rename all tests from percoa_slow_extended-X to X name
3) Extract start/stop/grep logic from tests to log_start/log_stop/log_grep includes
4) Remove tests mysql-test/slow_extended.patch/percona_slow_extended-use_global_long_query_time, because variable "use_global_long_query_time" was removed in 5.5 (and logic from this test) available in the slow_query_log_use_global_control.
5) Replace all SLEEP(X) by SET SESSION debug="+d,query_exec_time_X" or /* query_exec_time_X */
Also fixes bugs:
bug #794780 - "percona_slow_extended-microseconds_in_slow_extended sometimes fails"
bug #803071 - "add command-line variable setup test - log_slow_filter"
bug #803076 - "add command-line variable setup test - log_slow_slave_statements"
bug #803088 - "fix tests for log_slow_sp_statements:
bug #803093 - "add test for variable "log_slow_timestamp_every/slow_query_log_timestamp_always" (not command-line version)"
bug #803100 - "fix tests for use_global_log_slow_control/slow_query_log_use_global_control"
bug #803110 - "fix tests for slow_query_log_microseconds_timestamp/slow_query_log_timestamp_precision"
bug #803120 - "log_slow_verbosity ommited command-line tests in 5.5 "
bug #803137 - "rename percona_slow_extended-slave_innodb_stats.test to percona_slow_extended-log_slow_verbosity-innodb.test"
bug #760085 - "main.percona_slow_extended-slave_innodb_stats fails sporadically in 5.1.56 "
bug #803140 - "percona_slow_extended-min_examined_row_limit.test incorrect"

137. By Oleg Tsarev on 2011-07-11

Fixes for bugs #712393 #803867
1) I move query_response_time_collect call from intermediate of log_slow_statement to begin.
Reason of bug #712393 - log_slow_statement exit before call query_response_time_collect
2) I move all DBUG_EXECUTE_IF (what used in percona_query_response_time.test) from response_time_distribution.patch to slow_extended.patch
This functionality requires in the slow_extended sporadic fails
3) If we set "debug" variable:
    SET [SESSION|GLOBAL] debug="+d,query_exec_time_debug"
we can setup exact query_exec_time through set "debug" variable:
    SET [SESSION|GLOBAL] debug="+d,query_exec_time_TIMES";
where TIMES is [0.31|0.32|0.33|0.34|0.35|0.36|0.37|0.38|0.39|0.4|0.5|1.1|1.2|1.3|1.4|1.5|2.1|2.3|2.5]
After this commit we can also add special comment to the begin of query:
    /* query_exec_time="TIMES" */ QUERY;
This changes requires for test SQL_Slave_Thread query execution.
4) I replace logic:
/*----------------------------------------*/
    DBUG_EXECUTE_IF("query_exec_time_debug",
                    if (query_execution_time == 0)
                      opt_query_response_time_stats= 0;
                    else
                      opt_query_response_time_stats= 1;);
/*----------------------------------------*/
By following:
/*----------------------------------------*/
void query_response_time_collect_with_check(ulonglong query_exec_time)
{
  DBUG_ENTER("query_response_time_collect_with_check");
  bool collect= opt_query_response_time_stats;
  DBUG_EXECUTE_IF("query_exec_time_debug",
                  {
                    if (0 == query_exec_time)
                    {
                      collect= false;
                    }
                  });
  if(collect)
  {
    query_response_time_collect(query_exec_time);
  }
  DBUG_VOID_RETURN;
}
/*----------------------------------------*/
Otherwise debug-code has bad side-effect - set QUERY_RESPONSE_TIME_STATS to OFF if query shouldn't collect
5) I rewrite tests for response_time_distribution.patch, now all test unified and wouldn't have sporadic fails in the feature.
This change fixes the bug #803867.
6) I also remove mysql-test/response_time_distribution.patch and move to patch.
Now this is the suite "percona" and tests "query_response_time[-[stored|replication]]"
7) I also move have_response_time_distribution.[require|inc] to patch
8) I also add "percona_suite.patch" for include "percona" test suite to default suites list

136. By Stewart Smith on 2011-06-30

merge fdatasync() usage for innodb log files

135. By Stewart Smith on 2011-06-28

merge bug 802829 fix

134. By kinoyasu <kinoyasu@gauntlet4> on 2011-06-28

fix bug802348

133. By Stewart Smith on 2011-06-24

merge fix of valgrind errors in various federated test cases on 32bit valgrind.

132. By Stewart Smith on 2011-06-24

merge fix of hitting assert in 32bit debug builds

131. By Stewart Smith on 2011-06-23

merge fix fo percona_innodb_deadlock_count: now same test in 5.1 and 5.5

130. By Stewart Smith on 2011-06-23

merge fix of directory names: 5.1 and 5.5 PS are now the same as each other and we have nice symlinks to make build script better

129. By Stewart Smith on 2011-06-23

merge percona_server_variables fix

Branch metadata

Branch format:
Branch format 7
Repository format:
Bazaar repository format 2a (needs bzr 1.16 or later)
Stacked on:
lp:percona-server/5.5
This branch contains Public information 
Everyone can see this information.

Subscribers