[OpenAFS-devel] Re: AFS ... or equivalent ...

[Robert Watson]

On Sun, 13 Jan 2008, Jason C. Wells wrote:

> Marc G. Fournier wrote:
>
>> Does anyone know if there is any serious work being done to get AFS working 
>> under FreeBSD?  I have a large project that I'm working on that AFS (or 
>> something equivalent) would be *very* useful for, but we're trying to keep 
>> it as FreeBSD-pure as possible ...
>
> Yes.  Please get in touch with any of the people CC'ed in this list.  I 
> believe Matt Benjamin is the one who is actually getting serious on this 
> project.  Patches were even mentioned in a recent email.  I recall Jim Rees 
> is knowledgeable on AFS.  I also think one Derrick J. Brashear was 
> interested/knowledgeable too, but I don't have his address handy.  If I 
> misrepesented anyone please feel free to correct me.
>
> Matt, if you do not know Marc, look up Postgresql.  Marc is the port 
> maintainer for postgresql as well as a postgres developer. (iirc)
>
> Me, I am just a user who put together an ugly, ugly little FreeBSD port a 
> long time ago in the hope that it would inspire some people who were 
> qualified to do real work to pick it up and run with it.
>
> There are a couple mailing lists suitable for FreeBSD porting discussions. 
> One is run by the OpenAFS people and the other is run by FreeBSD people.
>
> Sorry for the spam and cross posts. It seems like the interest in OpenAFS on 
> FreeBSD is building.  I hope that this message will put the right people in 
> touch with each other and that maybe a concerted effort to port OpenAFS to 
> FreeBSD will arise.

Arla, which is just an AFS client, runs on some versions of FreeBSD, although 
typically not really recent ones.  I spent a little time this summer looking 
at getting it updated to 7, but ran out of time.

I'd like very much to get at least the kernel parts of an AFS client into the 
base system, as otherwise any AFS port (be it Arla, OpenAFS, etc) will 
constantly be falling behind and breaking as the base tree moves forward. 
Our VFS tends to change with moderate speed, and having it in the base tree 
will allow it to be updated as part of regular changes to our KPI by the 
author of the changes, rather than watching more and more ifdefs appear in a 
third-party tree.  I'm happy to lend a hand with this, but I don't have the 
time (apparently) to drive a port forward myself right now.

Robert N M Watson
Computer Laboratory
University of Cambridge