[Gluster-users] Pre-populating volumes

[Israel Shirk]

Hey guys, I'm a server admin working on some high-availability web server
setups with Amazon's EC2.  We're running Gluster to keep our assets sync'd
up between regions globally using replicated volumes with one brick in each
geographic region.  Each brick is a single EBS drive, so it's easy to
snapshot and create a new volume from the snapshot, taking maybe 30 seconds
total for around 100GB assets.

We've got a nice streamlined setup process for all our servers to keep
instantiations to a minimum latency (several minutes), with one notable
exception - getting our gluster assets volumes up to date always runs a bit
slow.  I've been trying to find some information on pre-populating the
volumes - ideally we'd like to just use the EBS snapshots mentioned above.
 While it seems fairly undocumented I've found enough references to it that
it seems worth looking into.  Does anyone know of any documentation for
this if it exists, or what caveats I may run into when running a setup like
this?

Thanks for the help,
Israel
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://supercolony.gluster.org/pipermail/gluster-users/attachments/20120801/a84d5bea/attachment.html>