[OpenAFS] OpenAFS with RAID

Horst Birthelmer horst@riback.net
Wed, 28 Dec 2005 15:07:39 +0100

On Dec 28, 2005, at 2:37 PM, Paul Robins wrote:
> I'll reply in one if that's ok (sorry for the topposting)
No problem ... :-)

> I would expect a disk to be the thing to go to be honest but  
> regardless, i want some system where there is parity data stored on  
> other nodes in this group of machines. Basically RAID5 but  
> networked would be perfect, as that would give me ~ 400 gig of  
> space whilst being able to handle a machine vanishing from the  
> network (the whole machine dies when the disk does, cheap  
> whiteboxes don't you know)


> I understand the way AFS works with regards to clients seeing /afs,  
> and i did see read only replication, and then running a command to  
> change a read only node(?) into a read write node (i'm sorry if i'm  
> talking crap, i'd read the wiki if i could). This is why i figured  
> perhaps it could be implimented with some sort of networked RAID5,  
> giving me a lot more storage than just RO mirroring one server to  
> the other 3, but whilst still being redundant.

You're not talking nonsense at all. It's exactly the kind of  
statement I was provoking.
(Just the term 'node' is from cluster terminology not AFS, but OK.
AFS doesn't care about nodes, there are only volumes on fileservers) :-)

I don't really recommend that conversion of RO volumes to RWs for  
backups. If you search the archives for that topic you'll find some  
of my old statements regarding that.
IIRC I said something like: use that only if your disk is gone, like  
in 'the dog eat your harddisk'. (I didn't look it up ... ;-) )

Well, reading your idea, I think I have to repeat myself.
It's doable, especially with something like ENBD, but you have to  
'pay' for that with some performance, as far as I can oversee that  
You'll have one fileserver then (or maybe 2) and transfer a lot of  
data over the network to those fileserver(s), just to transfer that  
from there to your AFS clients, in the worst case over the same network.

It _could_ work, even if it's not really what the designers of AFS  
had in mind when they built the system. ;-)