replicated file system?

[Mark Komarinski]

On 02/29/2012 10:22 AM, Kenny Lussier wrote:
>
>
> On Tue, Feb 28, 2012 at 3:36 PM, Michael ODonnell 
> <michael.odonnell at comcast.net <mailto:michael.odonnell at comcast.net>> 
> wrote:
>
>
>
>
>     > (DRBD>LVM>iSCSI>Heartbeat)
>
>     Heh.  I suspect that will somehow look familiar to Mr. Lussier... ;->
>
>     He did indicate a wish to have an Active/Active rig but I believe
>     that approach only allows Active/Standby, yes?
>
>
> I think I have seen that somewhere before :-)
>
> The issue with DRBD is that is is exactly like a network-based RAID 1. 
> The data written to the second server is not accessible unless there 
> is a failure on the first system. What I would really like to do is 
> have two (or more) servers share their local  file systems with each 
> other and load balance FTP and sFTP. If one server goes down, the 
> files that were written to it should still be available from other 
> nodes. Something like a K-safe file system. This would all be a lot 
> easier if there were a shared storage device in the background, or 
> even an NFS server.

As pointed out before, DRBD can do active/active, so long as the 
filesystem on top support it (GFS2 and OCFS2).  The DRBD team even has 
documentation to get you started:

http://www.drbd.org/users-guide/ch-gfs.html
http://www.drbd.org/users-guide/ch-ocfs2.html