[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

timewarp to 2052 (potato on alpha)



I just installed potato on a fresh-out-of-the-box DS20E.

IIRC, when it asked me whether or not the system time is kept in UTC
(and I answered affirmatively) it was displaying the correct date and
time (including the year, 2000).  That was amusing.  The system
hwclock time was correctly set to UTC as it came out of the box (or
does the boot floppy go and set the time off of the net somewhere).

However, later on (after a few reboots where I was messing with the
SRM stuff to make this thing auto-boot Debian by clearing boot_file
and setting boot_flags to the digit 0 so that aboot will choose the 0
option in /etc/aboot.conf), I noticed the date was wrong.  It was in
the year 2052.

I just checked on another DP264 machine on which I installed a fresh copy
of debian potato a couple of weeks ago, and it was in the year 2048!
(That machine was running red-hat 6.0 before I installed potato.)

I did this:

 sudo date 052617022000
 sudo hwclock --systohc
 sudo rm /etc/adjtime
 sudo shutdown -r now

on both machines.   That seems to have fixed it for now.


Why doesn't a fresh install of debian potato get this right?
Or did I do something wrong when I did the install?
(I do not recall ever being given the opportunity to set the time
though I do remember it showing me the time when it asked whether or
not the system time was to be kept in UTC.)

I have another pristine DS20E still in its cardboard box which I will
likely install Debian potato on within the next week.    Let me know
if you want any experiments done (like somehow reading the hwclock
before anything else has had a chance to touch it).

(I did once take a peek at the code in the kernel that tries to pick
an epoch.  Bleh.  So I do realize that this is an ongoing complication
of linux on alpha.  I'm suspicous that the problem got worse after
January of this year because one of the magic epoch-choosing windows
flipped into a different window, but it has been a few months since I
peeked at that code.)

			-Tim Shepard
			 shep@lcs.mit.edu



Reply to: