Conference smurf::buildhelp

2197.0. "bco hangs when sandbox dir is NFS mounted from guru" by AOSG::FILTER (Automatic Posting Software - mail to flume::puck) Fri May 03 1996 15:57

Date Of Receipt: 	 5-APR-1996 20:11:15.03
From: 	SMURF::GURU::kucherov "sergei kucherov  05-Apr-1996 2008"
To: 	odehelp@dec:.zko.guru, admin@dec:.zko.guru
CC: 	kucherov@dec:.zko.guru, [email protected], metsch@dec:.zko.guru
Subj: 	bco hangs when sandbox dir is NFS mounted from guru

I have verified Evelyn's claim that bco hangs when you're in
a workon on machine X and the actual sandbox directory is on guru,
and the sandbox directory is NFS mounted onto machine X from guru.
I have verified this for the cases where X=music (V3.2C)
and X=soil (V3.2D-1). I put the sandbox dir under my home directory,
which is on guru.

My guess is that this is yet another instance of the "well known"
NFS hanging problem in V3.2C which I reported (and recently QAR'd)
wherein sending mail via dxmail will hang if the user's home
directory is on guru and dxmail is started on any other system
which has the home directory NFS mounted. Somewhat surprisingly,
dxmail hangs even on tlaser (V4.0).

Does anyone know if guru has all the V3.2C patches applied?
If so, does anyone know of any pointers to the fix for this problem
in V4.0 so that I can backport it immediately and make myself happy
(for dxmail) and Evelyn happy (for bco).
BTW this would be a pro-active backport
which is in either my or Locus's domain to do if only someone would
help point us to the fix in V4.0 (Platinum).

	Thanks,
	sergei


P.S. Another solution would be to upgrade guru to V4.0, and just
bypass the problem. I don't see why this upgrade should wait for
the long-anticipated hardware improvements to guru.
On the other hand, last week I reported serious problems on
tlaser (V4.0) making vi unusable and csh usable only after editing
your .cshrc. I have no great respect for V4.0, so the best solution
might still be to actually *fix* the NFS hang problem in V3.2C.