List:Commits« Previous MessageNext Message »
From:Luís Soares Date:December 9 2010 10:44am
Subject:Re: bzr commit into mysql-trunk branch (andrei.elkin:3210) Bug#58753
View as plain text  
Hi Andrei,

On 12/07/2010 07:17 PM, Andrei Elkin wrote:
> Luis, hello.
>
>> Hi Andrei,
>>    Please find my review comments below.
>>
>> STATUS
>> ------
>>
>>    Not approved.
>>
>> REQUIRED CHANGES
>> ----------------
>>   n/a
>>
>> REQUESTS
>> --------
>>
>>    R1. Please, enlighten me! Why is this a sporadic failure?
>
> I was not sure it was sporadic actually.
> Until now when have just tested locally, and indeed no signs from *mtr*
> about the suspicious warning:
>
>
>     ==============================================================================
>
>     TEST                                      RESULT   TIME (ms)
>     ------------------------------------------------------------
>
>     worker[1]  - 'localhost:13010' was not free
>     worker[1] Using MTR_BUILD_THREAD 302, with reserved ports 13020..13029
>     rpl.rpl_checksum 'mix'                   [ pass ]   1361
>     rpl.rpl_checksum 'row'                   [ skipped ]  Doesn't support
> --binlog-format='row'
>     rpl.rpl_checksum 'stmt'                  [ skipped ]  Doesn't support
> --binlog-format='statement'
>     ------------------------------------------------------------
>     The servers were restarted 0 times
>     Spent 1.361 of 18 seconds executing testcases
>
>     Completed: All 1 tests were successful.
>
>     2 tests were skipped, 0 by the test itself.
>
>
> However, the warning is *not* sporadic.
>
>     $ cat
> /home/andrei/MySQL/BZR/2a-23May/mysql-trunk/mysql-test/var/log/mysqld.1.err
>
>     Suspicious lines from
> /home/andrei/MySQL/BZR/2a-23May/mysql-trunk/mysql-test/var/log/mysqld.1.err
>     ...
>     101207 22:10:05 [Warning] Master is configured to log replication events with
> checksum, but will not send such events to slaves that cannot process them
>
>
> So I would think of mtr in this case that for some reason is sensitive
> enough to react as it happened to PB2.
>

Please, file an MTR bug report for this.

>>
>>    R2. Aren't all tests vulnerable when checksums are active in
>>        MTR? If they are, wouldn't it be better to place the
>>        warning in include/mtr_warnings.sql ?
>
> Not this type.
> The failure in rpl_checksum is due to simulation and can only happen in
> NM ->  OS replication.

OK. Good. Then we should be fine with your fix. Patch
approved.


However, I have one last request before pusing (not really
related to the bug, but since we're improving
rpl_checksum.test), can you please remove the TODOs in
the test case? It should be straightforward...

Regards,
Luís

> cheers,
>
> Andrei
>
>
>>
>> SUGGESTIONS
>> -----------
>>   n/a
>>
>> DETAILS
>> -------
>>   n/a
>>
>> On 12/06/2010 01:08 PM, Andrei Elkin wrote:
>>> #At file:///home/andrei/MySQL/BZR/2a-23May/mysql-trunk/ based on
> revid:alexander.nozdrin@stripped
>>>
>>>    3210 Andrei Elkin	2010-12-06
>>>         Bug #58753 rpl_checksum finds unsuppressed warning
>>>
>>>         Suppressed a warning allowed by the test's logics.
>>>        @ mysql-test/suite/rpl/r/rpl_checksum.result
>>>           results are updated.
>>>        @ mysql-test/suite/rpl/t/rpl_checksum.test
>>>           suppression is added.
>>>

Thread
bzr commit into mysql-trunk branch (andrei.elkin:3210) Bug#58753Andrei Elkin6 Dec
  • Re: bzr commit into mysql-trunk branch (andrei.elkin:3210) Bug#58753Luís Soares7 Dec
Re: bzr commit into mysql-trunk branch (andrei.elkin:3210) Bug#58753Luís Soares9 Dec