Date: Fri, 15 Mar 2013 20:52:00 -0400 (EDT) From: Rick Macklem <rmacklem@uoguelph.ca> To: Freddie Cash <fjwcash@gmail.com> Cc: Adrian Chadd <adrian@freebsd.org>, Alfred Perlstein <bright@mu.org>, Andre Oppermann <andre@freebsd.org>, FreeBSD-Current <freebsd-current@freebsd.org> Subject: Re: NewNFS vs. oldNFS for 10.0? Message-ID: <636921049.3962434.1363395120474.JavaMail.root@erie.cs.uoguelph.ca> In-Reply-To: <CAOjFWZ4_XvXAQ4UsMX5=bQ%2BuSUkeNqNvnwd%2BrjhV7yYq-91yfA@mail.gmail.com>
next in thread | previous in thread | raw e-mail | index | archive | help
Freddie Cash wrote: > Isn't the general process (or at least past pattern) to: > - have 1 release cycle with just the old code (aka 8.x with oldNFS) > - have 1 release cycle with old and new code, default to old (aka 9.x > with oldNFS + newNFS) Actually, your numbering is out by one. 7.x - old only 8.x - old and new, with old default 9.x - old and new, with new default I, personally, don't care if 10.x has both in it or not. The only downside I see is that it often means generating patches for both, but this usually occurs when the code in the old and new are virtually identical (new cloned from old). rick > - have 1 release cycle with old and new code, default to new (aka 10.x > with newNFS) > > - remove the old code from next release (aka 11.0) > > > > Or is that too long of a time-frame to migrate from old to new? > > > > > > On Fri, Mar 15, 2013 at 11:46 AM, Adrian Chadd < adrian@freebsd.org > > wrote: > > > On 15 March 2013 11:11, Alfred Perlstein < bright@mu.org > wrote: > > > People in my org have been working with NFS and reporting issues for > > the > > past year. I'm quite certain that Doug White has reported issues due > > to > > missing certain caching features of the old code. > > > > This is not indicative that newNFS is bad, just that it still needs > > some > > work. > > Good news. and yes, it needs more work, but it doesn't preclude it > from having a cutover date set. Even if that date is something far in > the future, like 11.0. > > Or we'll just end up with two NFS stacks for some undetermined amount > of time. > > > Sure, and how much NFS do you actually use and support exactly? > > .. and exactly how much would that lend to this discussion? > > I'm not arguing NFS technical details, I'm arguing project forward > thinking and planning. These don't need me to be waist deep in NFS, it > needs a broader view of how things may and may not go. > > I lived through the pain of Linux having multiple NFS implementations > for precisely this reason. It was a clusterfsck of a nightmare of epic > proportions. We should avoid that. > > > > adrian > _______________________________________________ > freebsd-current@freebsd.org mailing list > http://lists.freebsd.org/mailman/listinfo/freebsd-current > To unsubscribe, send any mail to " > freebsd-current-unsubscribe@freebsd.org " > > > > -- > Freddie Cash > fjwcash@gmail.com
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?636921049.3962434.1363395120474.JavaMail.root>