[2754] in testers
Re: [daemon@ATHENA.MIT.EDU : sun4 7.7K: afsd]
daemon@ATHENA.MIT.EDU (Jonathon Weiss)
Wed Jan 4 17:20:44 1995
To: mhbraun@MIT.EDU
Cc: testers@MIT.EDU
In-Reply-To: Your message of "Wed, 04 Jan 1995 16:58:27 EST."
<199501042158.QAA07985@medic.MIT.EDU>
Date: Wed, 04 Jan 1995 17:20:21 EST
From: Jonathon Weiss <jweiss@MIT.EDU>
>> 1) Three sparc 5's in the w20 cluster (26,27,28) had their time
>> set 10 minutes slow at the time of installation. They presumably
>> lost this time while they were off, reflecting shoddy Sun hardware,
>> being a hardware bug. Complain to your SMCC sales rep.
>>
>> 2) The people installing the machine did not attempt to login and
>> confirm the machine worked (per last), nor did they reset the time.
These aren't even remotely release issues
>>
>> 3) /etc/init.d/athena attempts to run /etc/gettime instead of
>> /etc/athena/gettime to set the initial time upon boot. This doesn't
>> work as there is no /etc/gettime.
We probably ought to get this fix into the release, I'm surprised we
didn't notice it sooner.
>> 3a) gettime by default attempts to set time from a nonexistant .arpa
>> server. This should be fixed in the sources (I recommend time.mit.edu).
This may not be worth worrying about for the patch release.
>> *4*) afsd does not seem to correctly set the time. It should (since it
>> is not invoked with -nosettime) attempt to do so periodically,
>> and syslog it's attempts. I'm not familiar enough with afs
>> internals to do debug this problem, and the lack of a working
>> syslogd excacerbates my difficulties.
Works fine on my Sparc Classic...
>> 5) Comments in /etc/athena/rc.conf imply that setting TIMECLIENT to
>> true will activate timed:
>>
>> TIMECLIENT=true; export TIMECLIENT # Timed client?
>>
>> The comment should be either "# Time client", or TIMECLIENT
>> should cause timed to be run.
Yeah, I suppose we could at least correct this on the packs, not a
tremendous deal tho.
Jonathon
