[3927] in Release_7.7_team
Status of Solaris 9.2 public release
daemon@ATHENA.MIT.EDU (Garry Zacheiss)
Tue Jul 15 08:01:26 2003
Message-Id: <200307151201.h6FC1OH0014410@bart-savagewood.mit.edu>
To: release-team@MIT.EDU, ops@MIT.EDU
Date: Tue, 15 Jul 2003 08:01:24 -0400
From: Garry Zacheiss <zacheiss@MIT.EDU>
The public release of 9.2 for Solaris tonight was mostly uneventful.
I've only noticed two issues, neither cause for immediate concern:
- The update bug we encountered last year with USB zip drives is still
present: A Sun with a USB drive will see its internal IDE disk as
c1t0d0 when booted into the miniroot, but c0t0d0 before and after the
update, causing the miniroot boot to fail mounting / because the
generated vfstab refers to a nonexistant device.
This affected 5 machines, which I fixed by hand via the following
procedure:
mount -o rw,remount /dev/dsk/c1t0d0s1 /
ed /etc/vfstab
and change all occurances of c0 to c1, then reboot and let the update
happen.
We should resolve this one of these years, although doing so by
throwing the zip drives out the window may be a valid strategy.
- By around 4am, it was clear that the update of machines in w20-575 was
becoming bogged down, with much longer update times than machines in
other clusters. As of the time of this writing, there is still a
respectable number of Suns updating in w20.
I observed during the morning that w20-575a-rep-7 was bouncing on the
noc, and pinging it gave:
----w20-575a-rep-7 PING Statistics----
225 packets transmitted, 89 packets received, 60% packet loss
round-trip (ms) min/avg/max = 5/103/429
although whether this is the cause of the problem or a symptom of
pushing too much data through the network, I'm not sure.
As of around 5am, we were pushing about 98 Mb/s out of the w20 machine
room subnet, effectively saturating it. As of now, we're down to
about 43 Mb/s, so things are slowly tapering off on their own.
I don't think there's much to be done but wait for the machines to
finish updating, although we may wish to inquire as to whether the
repeater is a cause of slowness or just a victim of heavy bandwidth
utilization.
Garry
