MySQL Lists are EOL. Please join:

List:Commits« Previous MessageNext Message »
From:Sven Sandberg Date:September 28 2009 12:04pm
Subject:Re: bzr commit into mysql-pe branch (alfranio.correia:3514) WL#2687
WL#5072
View as plain text  
Alfranio Correia wrote:
> Hi Sven and Jasonh,
> 
> Alfranio Correia wrote:
>> Hi Sven,
>>
>> Thank you for the great review.
>>
>> I've addressed all your requests.
>> The only thing left is what follows:
>>
>>
>> Alfranio Correia wrote:
>> [...]
>>   
>>>>> --echo #13.e) Generates in the binlog what follows if a N-Table is
>>>>> changed:
>>>>> --echo #        --> STMT  "M B T R" entries if in M only N-Table
> is
>>>>> changed, format S.
>>>>>         
>>>> (S6) I think "B Nte T R" should be logged as just "Nte". We should
>>>> never write "B T R" in the binlog.
>>>>       
>>> This would be the ideal solution however we need to flush the
>>> transactional cache as the variable
>>> update_non_transactional_table is true and the cache's content is not
>>> empty and there is no way to
>>> check that the transactional cache does not hold any changes on
>>> non-transactional tables.
>>>     
>> I will be addressing this tomorrow while changing the server code.
>>   
> Sven, unfortunately this is not possible now. The problem stems from the
> fact that we
> have a single variable (i.e. transaction.stmt.modified_non_trans_table)
> to identify
> when a non-transactional table is changed. So if an M-statement,
> specifically,
> TN-statement, happens early in a the
> transaction.all.modified_non_trans_table is not
> being set appropriately in some situations. I think there is only one
> case in what this is not
> being set correctly, however, I need to investigate it more carefully.
> So I will later on file a bug
> report on this, fix it, and  then fix what you are suggesting in here.
> 
> Do you agree?

Yes, that sounds good. Please let me know the bug number when you have 
filed it!

/Sven

> 
> Cheers.
> 
>> Cheers.
>>
>>   
> 


-- 
Sven Sandberg, Software Engineer
MySQL AB, www.mysql.com
Thread
bzr commit into mysql-pe branch (alfranio.correia:3514) WL#2687 WL#5072Alfranio Correia13 Sep
Re: bzr commit into mysql-pe branch (alfranio.correia:3514) WL#2687WL#5072Sven Sandberg16 Sep
  • Re: bzr commit into mysql-pe branch (alfranio.correia:3514) WL#2687WL#5072Alfranio Correia17 Sep
  • Re: bzr commit into mysql-pe branch (alfranio.correia:3514) WL#2687WL#5072Alfranio Correia20 Sep
    • Re: bzr commit into mysql-pe branch (alfranio.correia:3514) WL#2687WL#5072Sven Sandberg21 Sep
      • Re: bzr commit into mysql-pe branch (alfranio.correia:3514) WL#2687WL#5072Alfranio Correia21 Sep
        • Re: bzr commit into mysql-pe branch (alfranio.correia:3514) WL#2687WL#5072Sven Sandberg23 Sep
          • Re: bzr commit into mysql-pe branch (alfranio.correia:3514) WL#2687WL#5072Alfranio Correia25 Sep
            • Re: bzr commit into mysql-pe branch (alfranio.correia:3514) WL#2687WL#5072Alfranio Correia28 Sep
              • Re: bzr commit into mysql-pe branch (alfranio.correia:3514) WL#2687WL#5072Sven Sandberg28 Sep