Skip site navigation (1)Skip section navigation (2)
Date:      Thu, 21 Jan 2010 08:27:35 +0100
From:      Ruben de Groot <mail25@bzerk.org>
To:        Steve Bertrand <steve@ibctech.ca>
Cc:        "freebsd-questions@freebsd.org Questions -" <freebsd-questions@freebsd.org>
Subject:   Re: High availability SQL server setup
Message-ID:  <20100121072735.GA22231@ei.bzerk.org>
In-Reply-To: <4B57A887.2030801@ibctech.ca>
References:  <4B57A887.2030801@ibctech.ca>

Next in thread | Previous in thread | Raw E-Mail | Index | Archive | Help

Here's an article about multi-master replication on MySQL:

http://onlamp.com/onlamp/2006/04/20/advanced-mysql-replication.html

It's not rocket science ;)

Ruben

On Wed, Jan 20, 2010 at 08:06:15PM -0500, Steve Bertrand typed:
> Hi all,
> 
> For some time, I've been considering consolidating all/most of our SQL
> databases (all MySQL) onto a single dedicated cluster setup.
> 
> I'm looking for feedback on the best way to do this.
> 
> All of the options I've considered so far have both their drawbacks and
> benefits. From what I can tell, there's no one single way to be able to
> have everything that I want.
> 
> Off the bat, I haven't found a way to create a cluster that can have
> more than one host in the cluster writable.
> 
> My objective would be to start with two very high end boxes. One would
> sit in my primary location, the other a few blocks away over a gi fibre
> link.
> 
> I would want the remote box to pick up immediately if the master server
> fails. I figure I could achieve this using network trickery for IP
> failover, CARP or the like and span a couple of vlans across the fibre.
> 
> I would want each SQL server connected to separate edge routers to
> ensure both server and network resilience. Each box has two GigE NICs,
> so off the bat, I'd have each box doing VRRP to two separate edge gear
> at each location.
> 
> My concern is, is that I can't envision how both boxes could possibly
> stay in a continuous state that would allow such fail-over, and
> fail-back. (fail-back is less of a concern...if it comes to it, I'd
> rebuild by hand if necessary).
> 
> I've considered ZFS replication, but there could be several minutes
> worth of snapshot missing if the primary fails.
> 
> I already have MySQL replication in many spots, but that's only one
> write master and read-only slaves.
> 
> Can you provide any details or new ideas that I'm missing in order to
> have the holy grail of SQL redundancy?
> 
> Cheers,
> 
> Steve
> _______________________________________________
> freebsd-questions@freebsd.org mailing list
> http://lists.freebsd.org/mailman/listinfo/freebsd-questions
> To unsubscribe, send any mail to "freebsd-questions-unsubscribe@freebsd.org"



Want to link to this message? Use this URL: <http://docs.FreeBSD.org/cgi/mid.cgi?20100121072735.GA22231>