Re: bzr commit into mysql-5.0 branch (skozlov:2613) Bug#22442

[Serge Kozlov]

I've updated test cases by your suggestions
http://lists.mysql.com/commits/52474

Andrei Elkin wrote:
> Serge,
> 
>> Andrei Elkin wrote:
>>> Serge, hello.
>>>
>>> Could you please start with the problem desciption?
>>>
>>> I think it is
>>>
>>> Problem: using sleep() in the test cased its indeterministic execution.
>>>
>>>> #At file:///home/ksm/sun/repo/mysql-5.0-rpl-bug22442/
>>>>
>>>>  2613 Serge Kozlov	2008-08-25
>>>>       Bug#22442:
>>>>       1. Added vertical output for SHOW SLAVE STATUS and waiting
> starting/stopping slave instead sleep.
>>>>       2. Updated the result file.
>>> ok
>>>
>>>> modified:
>>>>   mysql-test/r/rpl_flush_log_loop.result
>>>>   mysql-test/t/rpl_flush_log_loop.test
>>>>
>>>> === modified file 'mysql-test/r/rpl_flush_log_loop.result'
>>>> --- a/mysql-test/r/rpl_flush_log_loop.result	2007-10-10 07:21:11 +0000
>>>> +++ b/mysql-test/r/rpl_flush_log_loop.result	2008-08-25 09:23:15 +0000
>>>> @@ -20,5 +20,36 @@ master_password='',master_port=SLAVE_POR
>>>>  start slave;
>>>>  flush logs;
>>>>  show slave status;
>>>>
> -Slave_IO_State	Master_Host	Master_User	Master_Port	Connect_Retry	Master_Log_File	Read_Master_Log_Pos	Relay_Log_File	Relay_Log_Pos	Relay_Master_Log_File	Slave_IO_Running	Slave_SQL_Running	Replicate_Do_DB	Replicate_Ignore_DB	Replicate_Do_Table	Replicate_Ignore_Table	Replicate_Wild_Do_Table	Replicate_Wild_Ignore_Table	Last_Errno	Last_Error	Skip_Counter	Exec_Master_Log_Pos	Relay_Log_Space	Until_Condition	Until_Log_File	Until_Log_Pos	Master_SSL_Allowed	Master_SSL_CA_File	Master_SSL_CA_Path	Master_SSL_Cert	Master_SSL_Cipher	Master_SSL_Key	Seconds_Behind_Master
>>>>
> -#	127.0.0.1	root	SLAVE_PORT	60	slave-bin.000001	208	#	#	slave-bin.000001	Yes	Yes							0		0	208	#	None		0	No						#
>>>> +Slave_IO_State	#
>>>> +Master_Host	127.0.0.1
>>>> +Master_User	root
>>>> +Master_Port	9306
>>>> +Connect_Retry	1
>>>> +Master_Log_File	master-bin.000001
>>>> +Read_Master_Log_Pos	98
>>>> +Relay_Log_File	#
>>>> +Relay_Log_Pos	#
>>>> +Relay_Master_Log_File	master-bin.000001
>>>> +Slave_IO_Running	Yes
>>>> +Slave_SQL_Running	Yes
>>>> +Replicate_Do_DB	
>>>> +Replicate_Ignore_DB	
>>>> +Replicate_Do_Table	
>>>> +Replicate_Ignore_Table	#
>>>> +Replicate_Wild_Do_Table	
>>>> +Replicate_Wild_Ignore_Table	
>>>> +Last_Errno	0
>>>> +Last_Error	
>>>> +Skip_Counter	0
>>>> +Exec_Master_Log_Pos	98
>>>> +Relay_Log_Space	#
>>>> +Until_Condition	None
>>>> +Until_Log_File	
>>>> +Until_Log_Pos	0
>>>> +Master_SSL_Allowed	No
>>>> +Master_SSL_CA_File	
>>>> +Master_SSL_CA_Path	
>>>> +Master_SSL_Cert	
>>>> +Master_SSL_Cipher	
>>>> +Master_SSL_Key	
>>>> +Seconds_Behind_Master	#
>>>>
>>>> === modified file 'mysql-test/t/rpl_flush_log_loop.test'
>>>> --- a/mysql-test/t/rpl_flush_log_loop.test	2007-10-10 07:21:11 +0000
>>>> +++ b/mysql-test/t/rpl_flush_log_loop.test	2008-08-25 09:23:15 +0000
>>>> @@ -8,20 +8,22 @@ show variables like 'relay_log%';
>>>>   connection slave;
>>>>  stop slave;
>> Am I correct that we don't need to wait stop and because in fact next
>> statement is start slave and we have issues here?
> 
> right, no problem to execute
> 
> stop slave; start slave
> 
> without waiting for the first stop.
> 
>>> There is no need for waiting at all. Notice, that when `stop slave'
>>> returns the status' are
>>>  Slave_IO_Running	NO
>>>  Slave_SQL_Running	NO
>>> by the algorithm of `stop slave'.
>>> Hence, wait_for_slave_to_stop is never needed if the only point is to
>>> ensure the running status' are NO.
>>>
>>>> +--source include/wait_for_slave_to_stop.inc
>>>>  --replace_result $MASTER_MYPORT MASTER_PORT
>>>>  eval change master to master_host='127.0.0.1',master_user='root',
>>>>   master_password='',master_port=$MASTER_MYPORT;
>>>>  start slave;
>>>> +--source include/wait_for_slave_to_start.inc
>>> This synchronization is unnecessary as well according to logics of
>>> the test. Notice, that there is a difference with `stop slave' case in
>>> that Slave_IO_Running value is uncertain by return from `start
>>> slave', Slave_SQL_Running is YES.
>>> The uncertainly is due to the fact that Slave_IO_Running is
>>> three-value set. Returning ack to the client happens when IO thread
>>> has started but not yet connected to the master side.
>>> And only upon connnecting Slave_IO_Running gets YES.
>>>
>>>>  connection master;
>>>>  --replace_result $SLAVE_MYPORT SLAVE_PORT
>>>>  eval change master to master_host='127.0.0.1',master_user='root',
>>>>   master_password='',master_port=$SLAVE_MYPORT;
>>>>  start slave;
>>>> -sleep 5;
>>>> +--source include/wait_for_slave_to_start.inc
>>> ok.
>>>
>>>>  flush logs;
>>>>  sleep 5;
>>> what shall we do with this sleep?
>> The good question ... seems here we wait flush logs
>>
> 
> I think it is to wait for rotation of the relay log as the following
> command is 
> 
> show slave status;
> 
> Hence, why not to wait for the expected value of the relay log file
> with  wait_for_slave_param ?
> 
> A
> 


-- 
Serge Kozlov, QA Developer
MySQL AB, Moscow, Russia, www.mysql.com
Office:

Are you MySQL certified?  www.mysql.com/certification