I have 2 networks separated by a WAN (25/25 fiber optic VPN). I would like to provide some files to both servers on networks via NFS (and Samba for the desktops, but that's on a per-site basis). some of these files are core to the operation of the servers, so I would like some amount of redundancy between the two networks so that if the VPN link goes down, operations are not interrupted. there is no need for the synchronization of sites to be "instant", although I would prefer them to be quick instead of some sort of hourly rysnc job. I'm also interested in scaling to 3-5 network sites at some point in the future, so growth is a consideration.

from my research so far, there seem to be a few options:

1) drbd for sync between sites and then NFS on each side

2) clustered NFS

3) homebrew solution using unionfs, rsync or similar

it seems like drbd is actually the simplest solution, which for some reason seems counter-intuitive to me. are there another solutions I haven't found in my searching?

link|improve this question

80% accept rate
This sounds like a great fit for AFS en.wikipedia.org/wiki/Andrew_File_System or the (still in beta) ceph ceph.newdream.net – polynomial Sep 3 '11 at 19:20
I've been tracking ceph for quite some time, but didn't think it applied in this instance. I'm probably going to have one server node per network. there's also a number of components that go into ceph, making it even more complicated to set up and maintain than drbd! – neoice Sep 4 '11 at 2:34
feedback

Know someone who can answer? Share a link to this question via email, Google+, Twitter, or Facebook.

Your Answer

 
or
required, but never shown

Browse other questions tagged or ask your own question.