>I have a regular UFS filesystem exported from an Alpha 5000/500 running 3.0b.
>I mount this on an UltraSparc 1. *All* the files report timestamps that are
>one hour earlier than they should be.
>Both machines are otherwise functioning properly w.r.t. the local timezone.
The problem was the wrong timezone on Solaris (although the Solaris
installation software insisted on using WET for France!). Thanks to the
following for spotting this:
Olle Eriksson <olle_at_cb.uu.se>
Serge Munhoven <munhoven_at_olive.msm.ulg.ac.be>
Simon Tardell <tardell_at_particle.kth.se>
My confusion was really caused by our NTP setting the wall-clock time to be
correct in WET and therefore 1 hour wrong in MET (the timezone that the
Solaris machines should have been in). Setting the Solaris timezone to MET
cured the NFS timestamps, but broke the wall clock! After changing to MET, I
had to run ntpdate manually to correct the system clock by 1 hour to put the
system clock back to where it should have been.
Ian
Received on Tue Oct 29 1996 - 17:31:15 NZDT