replicated file system?

[David Miller]

On Tue, Feb 28, 2012 at 2:44 PM, Kenny Lussier <klussier at gmail.com> wrote:

> Hi All,
>
> I am looking for new ideas on how to replicate file systems. I have a need
> for redundant ftp servers, which could either be active/standby or
> active/active, as there is a load balancer in front of them. Currently, we
> periodically rsync the directory over to the standby system. What I would
> like to do is have a mirrored/replicated/clustered file system so that both
> systems can be active at the same time, and the data is automagically
> available on either, even in the event that one server fails. The catch is
> that there is no back-end shared storage (no SAN, NFS, etc.). I thought
> about drbd, but that is active/backup only. Most other systems required
> shared storage. I'm looking at using incron/inotify or Unison, but I was
> curious to see how other people would creatively solve this problem. Ideas?
>

I would have a look at DBRD.  It was merged into the Linux kernel a couple
years back in 2009 in version 2.6.33 of the kernel.  So it should be pretty
easy to find support for it without having to deal with 3rd party
repositories as long as you're not suck using RedHat or a derivative as
even RHEL 6 is using 2.6.32.

You can think of it as a network RAID 1 block device.  You use it similar
to a MD (Software RAID) device.
--
David
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.gnhlug.org/mailman/private/gnhlug-discuss/attachments/20120228/9307cad3/attachment.html