Would it be possible for DNS records to be created in the form of
int-mysql-slave.medialation.net that could automatically round robin the
requests?  That would probably be the best since u can then just point all
your reads to that hostname.

Anyone think that this is a valid scalr feature request?

Akshay


2009/9/15 Frédéric Sidler <[email protected]>

> Pay attention with the version of mysql-proxy you are using.
> The problem you mentionned is not a scalr problem. This is a application
> logic problem. I suggest you should check why your slave are far behind the
> master. You should maybe opt for a bigger instance that is known to be
> faster.
>
> You should also think about the logic of your application to give some time
> for the slave to be up-to-date.
>
> I suggest you to visit mysql forum and usual web application forum for this
> purpose.
>
> Finally, I'm using mysql-proxy 0.6.0 for more than a year without any
> interruption.
>
>
> On Tue, Sep 15, 2009 at 11:55 PM, favo <[email protected]> wrote:
>
>>
>> Hi,
>>
>> I would like to ask everyone how they handle new mysql slaves which
>> just started with a fresh copy of last mysql-data when scaling up.
>>
>> How are you splitting the read/write requests to these slaves while
>> the slave is behind the master?
>>
>> I'm currently playing with mysql-proxy to split read/write requests
>> but it seems a bit unstable and it needs to be restarted if a slave
>> has been started which interrupts mysql access for a few seconds. But
>> my main problem is that it also forwards read requests to slaves which
>> are too far behind.
>>
>> Thanks for any comments!
>> Mario
>>
>>
>>
>
> >
>

--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups 
"scalr-discuss" group.
To post to this group, send email to [email protected]
To unsubscribe from this group, send email to 
[email protected]
For more options, visit this group at 
http://groups.google.com/group/scalr-discuss?hl=en
-~----------~----~----~----~------~----~------~--~---

Reply via email to