Skip site navigation (1)Skip section navigation (2)
Date:      Mon, 28 Nov 2005 14:27:50 +0100
From:      Ivan Voras <ivoras@fer.hr>
To:        mike <mike503@gmail.com>, hackers@freebsd.org
Subject:   Re: Possible way to distribute NFS?
Message-ID:  <438B05D6.3000108@fer.hr>
In-Reply-To: <bd9320b30511272005q43cb974dm47d5fbaa1e389f90@mail.gmail.com>
References:  <bd9320b30511271904l77cf6d3ar8b7eb50425c4cf78@mail.gmail.com> <bd9320b30511272005q43cb974dm47d5fbaa1e389f90@mail.gmail.com>

next in thread | previous in thread | raw e-mail | index | archive | help
mike wrote:

> If someone could write an NFS client using FUSE (perhaps?) and have it
> integrate with memcached[1], couldn't that basically allow for n+1 scaling

I have had a similar idea to do in FUSE for some time now, but I'm 
waiting until I can spare the time to do it and the FUSE system becomes 
more stable (which should happen soon).

Also note that, if done with FUSE, it will be slower than kernel code. 
Maybe not by much 
(http://creo.hu/pipermail/fuse4bsd-devel/2005-October/000009.html), but 
it's not yet clear how much.

The problem I would try to solve is that of having single-writer- 
multiple-readers setup (writes go through one machine, get distributed 
to other machines, reads can go wherever), and without using NFS (though 
it's a good idea now that you mention it :) ), so it's maybe not what 
you need.

If understand your suggestion correctly, all "NFS-like" traffic would 
pass through a single machine, which doesn't sound good speed-wise...?



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?438B05D6.3000108>