Re: [9fans] reliability and failing over.

[Eric Van Hensbergen <ericvh@gmail.com>]

On 9/9/05, Francisco Ballesteros <nemo@lsub.org> wrote:
> Funny. The 9p reliability project looks to me a lot like the redirfs
> that we played with before introducing the Plan B volumes into the kernel.
>

Yes, Gorka and I talked about this when he first started down the
path.  There were some differences that seemed important at the time,
but I can't recall what they were.
 
> 
> Probably, the increase in latency you are seeing is the one I'm going to
> see in volfs. The 2x penalty in performace is what one could expect, because
> you have twice the latency. However, the in-kernel implementation has no
> penalty at all, because the kernel can rewrite the mount tables.
> 
> Maybe we should talk about this.
> Eric? Russ? What do you say? Is it worth to pay the extra
> latency just to avoid a change (serious, I admit) in the kernel?
> 

Its something important to figure out.  I'll admit we are testing the
worst case scenerio here -- but the application we started the work
for (inter-partition communication) is very low-latency, and so its
desirable to eliminate as much overhead as possible.  Still, its worth
running local area network tests, but I'm going to go for gigabit
versus 100 mbit.  Gorka and I will work on this next week while we
finish up the recover paper.

Its quite likely I'll look at integrating something like this service
into my library-OS version of the 9P client and/or v9fs.  Still not
sure if it would make sense in Plan 9 proper or not.

        -eric