[1636] in Release_7.7_team
Re: 8.2.17/8.2.18 patch release
daemon@ATHENA.MIT.EDU (Thomas H. Grayson)
Wed Feb 3 19:37:52 1999
To: Mike Barker <mbarker@MIT.EDU>
Cc: f_l@MIT.EDU, network@MIT.EDU, release-team@MIT.EDU, 11.520staff@MIT.EDU,
phils@MIT.EDU, karen@MIT.EDU
In-Reply-To: Your message of "Wed, 03 Feb 1999 17:10:40 EST."
<4.1.19990203164822.00bd59a0@po8.mit.edu>
Date: Wed, 03 Feb 1999 19:37:34 EST
From: "Thomas H. Grayson" <thg@MIT.EDU>
Mike,
Thanks for your thoughtful reply to my request. A special version of the
'attach' program will work fine for our purposes. We typically ask users to
access our data using the commands 'setup 11.520', 'setup crldata', or 'setup
oracrl'. The '.attachrc' scripts in these lockers typically attach other
lockers and other setup tasks. It would be easy to include a '-f' parameter
to 'add' in these scripts, or to use the full path to a special version of
'attach'.
I have not done extensive benchmarking with the NFS rsize and wsize options,
so I don't know how much effect using custom parameters would have. All
things being equal, I'd like to be able to tweak these easily if something
comes up; we've experienced NFS timeout problems in the past that required
special intervention, and I'd like to have some tools to use if we experience
a new round of issues. On the other hand, I don't want to make things harder
or hold up the release. For now, I've seen excellent results by using the NFS
defaults. I'm happy to run with that if adding the ability to adjust the
parameters is too much work.
We could place the special attach program in the 'crlbin' locker, which would
ordinarily be the best place for special programs used at the CRL.
Thanks,
Tom Grayson
> Hello.
>
> Your request for a modification in the new Athena release was transferred to
> me as leader of the release team.
>
> At this time, we are not going to put the change you suggested into
> 8.2.17/8.2.18. Basically, the modification would affect all of the NFS
> users and the network, and there is insufficient time to test the impact of
> this modification.
>
> However, we are working on an alternative to help with your class lockers.
> Specifically, this is a "hacked" version of the current attach program. We
> have two questions: 1) is it sufficient to always use the system defaults
> for rsize and wsize or do you need to be able to adjust these? And 2) is
> there a location in your class locker(s) for a special program? Where?
>
> This solution will require that people either (a) add -f 11.520 (or whatever
> the locker name is) or (b) use a specific path to the special attach program.
>
> I realize that we are not taking the specific action you suggested. I
> think the alternative we have developed will provide you with the
> functionality you need of being able to attach NFS lockers using the system
> defaults for rsize and wsize. I hope you will agree to this expedited
> approach.
>
> Mike Barker
> MIT Information Systems