Today's testing at CSMC

[John Watlington]

We came in this morning to find that almost all machines not equipped  
with the
latest (.3) kernel had crashed with trac ticket #4322.  The  
exceptions to this
were one machine which had crashed w. #4431, one which had powered
itself off, and one which had rebooted.

We've upgraded all machines to the .3 kernel, and have not seen any more
crashes due to #4322.

So far, in our testing, we have seen two different machines power off  
on resume.
Another board has rebooted itself three times, and will be X-rayed  
tmw. to look
for assembly problems.   We have seen no other machines reboot.

We have seen another instance of #4431 (WLAN interface going away) on  
the
same machine, and eight instances of #4121 (WLAN interface doesn't go  
away,
but no longer works) over the course of the day.  We may well X-ray  
the #4431
machine as well.

One machine had its USB power switch die during the day.  We replaced  
the
chip and it's back in testing.

Seven of the new machines are in cases w. keyboard and display, and  
they all
refuse to start X (see previous email to devel).   Also troubling is  
a gray screen
of death on these machines after a kernel crash.   This feels like  
more of the
DCON regression that we thought was fixed in build 617 and later.

The serial dongles were released from customs, and arrived around 11AM.
We then wasted a large amount of time cabling them up only to realize  
that
the USB hubs purchased for us were not stable enough to use for logging!
We will purchase more tmw. morning, but may have to travel back to  
Shanghai
(90 minutes) to find a store stocking them.

Chris attempted to install the manufacturing server (we were going to  
use it
to log additional channels until we found more USB hubs), but was as  
confused
as the Ubuntu installer was by there being no hard disks inside it.   
Oops.

G'evening,
wad and Chris