maria-developers team mailing list archive

[Robert Hodges <robert.hodges@xxxxxxxxxxxxxx>]

Hi Seppo, 

Thanks, that was my assumption as well but life tends to a little more
complex than theory.

Cheers, Robert

On 1/30/10 1:10 AM PST, "seppo.jaakola@xxxxxxxxxxxxx"
<seppo.jaakola@xxxxxxxxxxxxx> wrote:

> Hi Robert,
> 
>>> Tungsten consistency checking technology works very well, and there
>>> is no need to "fix it" in any way. However, this method is not directly
>>> usable for multi master replication, because the target node(s) may
>>> have committed some transactions not yet seen in the originating
>>> master node, and these new transactions can interfere with the
>>> consistency check query.
>>> 
>> I'm sure I should really know this already, but does your statement amount
>> to saying the consistency check needs to execute in total order with other
>> change sets?  Or are there other data visibility issues along the lines of
>> non-repeatable reads that would cause you to get different answers on
>> different servers?
>> 
> 
> Consistency checking needs to be run in same read view in all the nodes.
> (this works only for MVCC engines).
> We can use total order monitor to control the start of the checking
> transactions, the actual check query can then be relaxed from total order.
> Which is actually crucial for this to work efficiently, as total order
> operations are very expensive.
> 
> Seppo
> 
> --
> http://www.codership.com  seppo.jaakola@xxxxxxxxxxxxx
> 
> 
>