[Discuss] Redundant array of inexpensive servers: clustering?

[richard.pieri at gmail.com (Richard Pieri)]

Rich Braun wrote:
> It's 2014 and I figured that maybe the state of the art in RAIS (true
> clustering of servers vs. disks) might have gotten somewhere since the last
> time I looked at the idea in about 2011.

There's no such thing. The only major implementation of paired 
(mirrored) processing that I'm aware of is Tandem's NonStop product now 
sold by HP. It's not inexpensive.

The next closest thing is, well, nothing, really. Nobody's done tight 
clustering with active process migration since VMScluster and 
TruCluster, a port of VMScluster to Tru64 UNIX.

Synchronous transaction processing like Galera isn't the same thing. Nor 
is live VM migration.

You can get something usable on the cheap with Xen: automate checkpoint 
file creation, use DRBD to replicate checkpoint data from hot to cold, 
use a heartbeat or manual switch to toggle nodes between hot and cold 
states. You'll still lose anything between the fault and the most recent 
checkpoint but that's what you get for cheap. Zero data loss and zero 
down time is costly.

-- 
Rich P.