[OpenAFS] Re: [OpenAFS-devel] Re: Thinking about 1.6

Derrick Brashear shadow@gmail.com
Wed, 16 Dec 2009 13:46:04 -0500

On Wed, Dec 16, 2009 at 1:38 PM, Andrew Deason <adeason@sinenomine.net> wrote:
> On Wed, 16 Dec 2009 18:04:58 +0000
> Simon Wilkinson <sxw@inf.ed.ac.uk> wrote:
>> *) Remove the --enable-bos-restricted switch, and associated #ifdefs
>> and make this behaviour the default - it's still controllable from the
>> command line, and the default case is safe.
>> *) Remove the --enable-supergroups switch, and enable supergroups by
>> default. A number of sites have been running with supergroups enabled
>> for years - making them optional just complicates configuration.
>> Before doing this we need to ensure that a downgrade tool is
>> available.
> For these two, we also need to make these really obvious and difficult
> to just pass by. Otherwise sites that depend on e.g. 'bos exec' would be
> in for a surprise.

bos exec still works unless you give the restricted command line
switch. if you turn on random options without reading what you're
doing, you get what you paid for.

>Or sites that assume they can just rollback to 1.4
> binaries and be fine (which they may not for supergroups).
>> *) Make demand attach the default, but provide
>> --disable-demand-attach-fs to allow old-style fileservers to still be
>> built
> I'm not sure I see any benefit in using old-style fileservers, but
> forcing people to change I guess may be going a bit far. Removing the
> DAFS ifdef greatly simplifies a lot of code, though. Would it be
> objectionable to remove it for the next dev branch?

in an environment where pthreads don't exist, i suspect DAFS would be
not that useful.

the question (to me) boils down to "how much do we care about that
environment" and i'd argue
until we no longer provide an LWP fileserver we de facto do. we should
explore that question, but it's out of scope for 1.6 regardless.

>> Comments?
> +1 on all other points. In addition, remove --enable-unix-sockets, which
> is already (confusingly) the default, and remove the associated #ifdefs.
> Is there any reason to not use unix sockets for SYNC?

no unix sockets on Windows and possibly some platforms depending on
kernel configuration.

> And not really a feature change, but perhaps make --enable-warnings the
> default when we have gcc?
> Are we / how long are we keeping the inode fileserver backend around?

for sites with solaris 8, might as well let them upgrade to 1.6.
anyone else, well, i hope they aren't still using it.