[3042] in Release_7.7_team
Pair of errors in 9.0.20 for Solaris
daemon@ATHENA.MIT.EDU (Greg Hudson)
Thu Nov 8 09:24:11 2001
Date: Thu, 8 Nov 2001 09:24:06 -0500
Message-Id: <200111081424.JAA10541@egyptian-gods.MIT.EDU>
From: Greg Hudson <ghudson@MIT.EDU>
To: release-team@mit.edu
I made a couple of mistakes in 9.0.20 which fortunately cancelled each
other out, for the most part. They were:
1. I forgot to release the system.sun4x_58.minirt volume in the dev
cell, which contains the oscheck stats. (Although I did include
it in the volume release list in my propagation instructions.)
The result is that the sparcv9 math library needed for cryptorand
didn't stay local on PUBLIC=true machines, so cryptorand didn't
start at boot time on those machines. (Which means ssh would
hang, since /dev/random wouldn't work.)
2. My propagation instructions did not successfully propagate the
packages in cdrom, because system/sun4x_58/install/cdrom is a
link to ../cdrom.
The result is that SUNWski didn't get installed on athena cell
machines.
Since there are very few PUBLIC=true dev cell machines, there was very
little lossage; mostly, we just failed to install SUNWski.
Problem 1 is something I have done at least once before (with the
netra packages). I'll make sure it doesn't happen in the future by
making a script for myself to release all the Solaris volumes.
Problem 2 shouldn't happen again now that I'm aware of the symlink.
Here is my plan: today, I will put out a 9.0.21 for Solaris in the dev
cell which reinstalls SUNWski and brings the sparcv9 math library
local again, and doesn't suffer from the above problems. We will put
it out in the athena cell next Monday.
