[2662] in Release_7.7_team

home help back first fref pref prev next nref lref last post

Re: [cfox@MIT.EDU : blank screen during update]

daemon@ATHENA.MIT.EDU (Garry Zacheiss)
Thu Mar 22 15:20:55 2001

Message-Id: <200103222020.PAA16535@big-gay-als-animal-sanctuary.mit.edu>
To: Greg Hudson <ghudson@MIT.EDU>
cc: release-team@MIT.EDU
In-Reply-To: Your message of "Thu, 22 Mar 2001 11:40:44 EST."
             <200103221640.LAA16266@nephthys.mit.edu> 
Date: Thu, 22 Mar 2001 15:20:46 -0500
From: Garry Zacheiss <zacheiss@MIT.EDU>

>> Normally, I would say, "okay, looks like we need to do a miniroot
>> update," since we have the machinery in place for that.  But it so
>> happens that miki discovered yesterday that the miniroot doesn't work
>> for machines already running Solaris 2.7 (afsd hangs, even if you clear
>> the cache first; it's not clear what afsd can see on a 2.7 machine
>> that's different from a 2.6 machine, so it's hard to know what to look
>> at).

   It would surprise me if this wasn't fixable.  Is the miniroot being
used here the same one we used for the 8.3 -> 8.4 upgrade?  What AFS
client is being used in it?  An upgrade to the 3.6p1 client might help.
Have we considered using a memcache for the miniroot afs client?  That
might also make a difference.   Has anyone tried running afsd with
-verbose and -debug?  That might be enlightening as well.

>> So, options:
>> 
>> 	* Put a loud warning in mail to release-announce.  Suboptimal
>> 	  because not everyone who watches machines auto-updating is
>> 	  on release-announce, and even people who are might not read
>> 	  the mail or might not read it or remember it.

   I don't think this is an option.  Even if we were able to convince
ourselves it's ok, Jeff would still make sure we got 30 years in the
electric chair for doing this.

>>	* Fix the problem.  Optimal, but could take a while.

   This is obviously the best solution.

>>	* Give up on trying to put out these Solaris patches.

   I think we need input from alexp before we can determine how much of
an option this is.  On the other hand, has anyone verified that the
patches in 8.4.21 make our JDK problems go away?  If they don't, punting
seems like the easiest solution.

Garry

home help back first fref pref prev next nref lref last post