> I whole heartily agree (for knfsd anyway, I don't use RAID): knfsd
> has been vaguely unstable for as long as I can remember and even now
> if someone asks me about NFS on Linux I can't tell them "sure, go ahead,
> it works fine". It's just one big mess with various incompatible patches
> floating around, noone knowing which patch subsumes which other and
> which patch works with which utilities.
>
> Compatibility is an issue once a service is stable, but right now, I
> bet that most NFS users wouldn't care about upgrading their tools since
> they already have to go through similar trouble just to get NFS to work
> properly in a heterogenous environment.
>
> It should be a priority to merge knfsd patches quickly in order to
> eliminate this big ugly mess and finally get to a state where sysadmin
> can use NFS with Linux without loosing their sleep.
I don't normally join in with "me too"'s, but knfsd is a sore subject with
me. I'll accept that Linus may not want Trond's nfs3 patches at this
point, but the stock client code is workable as-is for v2 (though somewhat
slower). The server is another story entirely.
HJ Lu has done some great work on achieving compatibility with non-linux
platforms and flushed many gremlins from the locking code. I haven't had
a single problem with his server patches and utilities (and, thanks to
Trond, they now coexist with his nfsv3 client patches <g>). Lest anyone
think I lead a charmed life, rest assured that if there is a problem with
a networked filesystem lurking, I will trip over it in a short time.
It's a royal pain having to apply these to every new kernel - and I'm
moderately experienced with such matters. For a newbie, the confusion
level must be off the scale.
Linus, Alan, can we please just take the step of putting HJ's patches into
the mainstream kernel, or failing that communicate clearly what you object
to? The current state of affairs is just not acceptable, IMHO.
Many thanks.
Steve
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/