Re: MySQL High Availability

[Frédéric VANNIÈRE <frederic@vanniere.fr>]

MySQL Cluster currently requires a huge amount of memory because the  
data
storage is done on memory. A disk storage may be available in the  
future.

Frédéric.


Le 9 nov. 05 à 22:23, Neil McGovern a écrit :

> On Wed, Nov 09, 2005 at 10:09:19PM +0100, Frédéric VANNIÈRE wrote:
> > Hi,
> >
> > I'm looking for a good solution for having MySQL redundancy.  There
> > are more than
> > 2000 databases and 7 GB of data.
> >
> > As now I've got two solutions:
> >    - using replication, it must be monitored and requires long
> > downtime for initialization and
> >      fallback.
> >    - using DRBD between a master and slave server, I'm using it for
> > a NFS server and it works
> >      pretty well. The problem is the slave may be corrupted
> > specially for InnoDB databases.
> >
> > There is also sqlrelay,  The best solution would be a 2 way
> > replication with load balancing and
> > caching (writes on 1 server, reads on several one)
>
> You may also want to look at MySQL Cluster, which is part if MySQL 5.0
> Max.
>
> Regards,
> Neil
> --
>    __
>  .`  `. neilm@debian.org | Application Manager
>  : :' ! ---------------- | Secure-Testing Team member
>  '. `-  gpg: B345BDD3    | Webapps Team member
>    `-   Please don't cc, I'm subscribed to the list