View Single Post

   
  #4 (permalink)  
Old 02-28-2008, 06:34 AM
Jimmy Guerrero
 
Posts: n/a
Default Re: MYSQL data replication

Hello,

Hmmm, I just reread your post and noticed the requirement of 50-60
slaves hanging off a single master. DRBD would NOT be the right solution
here.

-- Jimmy

Jimmy Guerrero wrote:
> Hello,
>
> Although you state that there is no requirement for near real time
> synchronization, an alternative might be to look into DRBD. Which if you
> are not familiar with, is block-level replication.
>
> See: http://www.mysql.com/products/enterprise/drbd.html
>
> Combining DRBD with Linux Heartbeat also gives you failover capabilities.
>
> Mohd posted a blog on his experience yesterday.
>
> See: http://blog.irwan.name/?p=118
>
> Jimmy Guerrero
> Sr Product Manager
> MySQL, Inc
> Houston, TX
>
> sol beach wrote:
>> I have limited experience with MYSQL replication; which is why I am
>> hoping
>> others with more experience can answer a question or two.
>> Let's say I have a MASTER MYSQL database.
>> Let's say there are 50 - 60 other systems where I'd like to have MYSQL
>> running on these "slave" systems.
>> These slave systems need to be kept in synch with the Master, but it does
>> NOT need to be anywhere near real time.
>> The data in the slaves could lag as much a an hour or two.
>> The amount of data in total in the MASTER is in the range of 100MB -
>> 250MB
>> The rate of changes to the data is in the range 2000 - 5000 DML per 24
>> hour
>> day.
>> We control the application so we can/will include date/time each
>> record is
>> created or modified.
>> You can assume that no records ever get physically deleted; only INSERT &
>> UPDATE (no DELETE).
>>
>> What are some alternative ways to keep the slave systems "current"?
>>
>> TIA & HAND!
>>

>

Reply With Quote