[GLLUG] Ubuntu Server Failover

Jean van Wyngaardt gabrieldrumming at gmail.com
Fri Aug 2 13:07:21 UTC 2013


Thanks John

Really interesting, I'll look at those versions of lsyncd and csync - I'll have to experiment quite a bit I guess.

Will be a fun project none-the-less!

J

On 2 Aug 2013, at 13:44, John Edwards <john at cornerstonelinux.co.uk> wrote:

> On Fri, Aug 02, 2013 at 10:54:38AM +0100, Jean van Wyngaardt wrote:
>> Hello Guys
>> 
>> Anyone have a good experience setting up a failover server, that also replicates the data from a master server?
>> 
>> I would assume rsync would do nicely for the data side, but I haven't really messed around too much with any failover apps before. Anyone have any favourites?
> 
> 
> Instead of rsync at certain intervals will give you a certain lag,
> you could look at either csync or lsyncd, which are both designed
> to quickly mirror changes made to one system onto another.
> 
> An alternative to that is to have both servers look at a common
> backend data store (iSCSI, NFS, etc).
> 
> 
> For failover (instead of clustering) the Linux-HA Heartbeat system
> can use will detect if the one machine is unavailable and tell the
> other to take over. Between two machines it's probably best to run
> it over a serial cable so you don't have to rely on the network
> being 100%, but I think it can be run over the network if serial is
> not possible.
> 
> Ucarp is a similar system that only uses the network and is a Linux
> implementation of OpenBSD's OpenCARP system, so could work across
> different operating systems.
> 
> 
> 
> -- 
> #---------------------------------------------------------#
> |    John Edwards   Email: john at cornerstonelinux.co.uk    |
> #---------------------------------------------------------#
> _______________________________________________
> GLLUG mailing list
> GLLUG at mailman.lug.org.uk
> https://mailman.lug.org.uk/mailman/listinfo/gllug





More information about the GLLUG mailing list