Distributed network RAID

[Bart Silverstrim]

Sean Carolan wrote:
>> For example, I would be able to take /home on machine 1 and any changes 
>> I make will automatically migrate over to machine 2's /home, then if 
>> machine 1 dies for some reason I can log into machine 2 and my data 
>> would be intact until I can bring up machine 1 again.
>>
>> Anyone else recall seeing this, or even better, configure something like 
>> this before?
> 
> Try Heartbeat (linux-ha) and drbd.  I have set this up and it works quite well 
> for creating a high-availability NFS share.

Aha!  I think this was the original implementation that I had heard about.

How robust is this setup in your experience?  Quirks?  Problems?

And how difficult is it to maintain or recover from a problem with this?

Thanks!