>
> I am a little out of my depth here so forgive me if I am taking rubbish,
> but I am not sure that having two servers constantly available is the right
> solution for you. Typically this setup would be used to spread the load if
> the application is in high demand. To provide high availability in the event
> of failure you would typically use some clustering software which will
> failover of IP addresses etc from the failed machine to another machine
> which sits idle ready to be used in the event of a failover.
I was thinking of that also. The LVS routers that have failover support
could actually be the DSpace instances then if one dies the other just trips
in. However I would still need to ensure that all the DSpace instances were
in sync, and for this I would need a common asset store and postgresql
instance.
I just did a quick google on 'Centos clustering' and this came up
> http://www.linuxtopia.org/online_books/centos_linux_guides/centos_cluster_configuration_and_management/.
> I don't know if its free or if there are any free alternatives.
This is the setup that I'm considering using.
>
>
> Cheers, Robin.
>
>
> Robin Taylor
> Main Library
> University of Edinburgh
> Tel. 0131 6515208
>
> > -----Original Message-----
> > From: John Preston [mailto:byhisde...@gmail.com]
> > Sent: 20 April 2009 00:16
> > To: dspace-tech@lists.sourceforge.net
> > Subject: [Dspace-tech] High Availability DSpace
> >
> > I have to set up a DSpace instance and want to make it highly
> > avalable. Its not that the load is gonna be that heavy to
> > start with, but I want to have some redundancy such that if a
> > machine goes bad, the system will still function while I look
> > about changing it out. So I'm planning on using LVS on Centos.
> >
> > I will have two routers which will keep traffic going to two
> > webservers on the backend which have duplicate DSpace
> > instances. Now to have duplicate instances I will use a
> > common NFS store for the assest store, but for the database I
> > will point each of the DSpace instances to a common database
> > on the network. I know this will be a single point of
> > failure, but I can either restore the database quickly if I
> > need, or I can get synchronisation across postgres databases
> > dynamically if I want (slony, etc).
> >
> > My question is: can anyone give me a simpler or better way to
> > achieve high availability, and is DSpace code able to manage
> > different web applications writing to the database across
> > different instances.
> >
> > John
> >
> >
> >
> >
>
>
> --
> The University of Edinburgh is a charitable body, registered in
> Scotland, with registration number SC005336.
>
>
------------------------------------------------------------------------------
Stay on top of everything new and different, both inside and
around Java (TM) technology - register by April 22, and save
$200 on the JavaOne (SM) conference, June 2-5, 2009, San Francisco.
300 plus technical and hands-on sessions. Register today.
Use priority code J9JMT32. http://p.sf.net/sfu/p
_______________________________________________
DSpace-tech mailing list
DSpace-tech@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/dspace-tech
