My alpha seems to be having major clock lossage (ntpd is stepping it
forward about .5 seconds every 20 minutes) and one of my i386 machines
seems to be rather unstable.

My ntp server is running NetBSD/i386 1.5, and is an old 386/33. The only
time steps ntp has reported since February are:
Feb  3 04:56:47 sloth ntpd[2495]: time reset -0.138986 s
Feb  3 06:08:26 sloth ntpd[2495]: time reset 0.151157 s
Feb  6 12:40:15 sloth ntpd[2495]: time reset 0.183325 s
Feb 16 19:19:53 sloth ntpd[2495]: time reset 0.145857 s
Feb 16 20:48:39 sloth ntpd[2495]: time reset -0.184374 s
Mar 19 03:09:50 sloth ntpd[2495]: time reset 0.166503 s
Mar 19 03:28:12 sloth ntpd[2495]: time reset 0.296566 s

My alpha is an AlphaPC 164 500MHz running NetBSD 1.5S from March
17. Just in the past few hours, it's done this:
Mar 23 22:03:49 yerfable ntpd[186]: time reset 0.448284 s
Mar 23 22:25:14 yerfable ntpd[186]: time reset 0.512497 s
Mar 23 22:46:40 yerfable ntpd[186]: time reset 0.538327 s
Mar 23 23:07:08 yerfable ntpd[186]: time reset 0.543180 s
Mar 23 23:28:31 yerfable ntpd[186]: time reset 0.492591 s
Mar 23 23:48:55 yerfable ntpd[186]: time reset 0.500918 s
Mar 24 00:10:26 yerfable ntpd[186]: time reset 0.498820 s
Mar 24 00:30:56 yerfable ntpd[186]: time reset 0.521543 s
Mar 24 00:51:29 yerfable ntpd[186]: time reset 0.523078 s
Mar 24 01:11:59 yerfable ntpd[186]: time reset 0.555755 s

And the i386 is a AMD K6-2 450MHz running NetBSD 1.5S:
Mar 19 03:46:49 dahan ntpd[224]: time reset 0.300060 s
Mar 19 06:57:25 dahan ntpd[224]: time reset -0.322411 s
Mar 22 00:40:35 dahan ntpd[224]: time reset -0.165401 s
Mar 22 01:20:23 dahan ntpd[224]: time reset 0.218928 s
Mar 22 01:40:49 dahan ntpd[224]: time reset 0.230840 s
Mar 22 02:01:21 dahan ntpd[224]: time reset 0.184014 s
Mar 22 06:55:28 dahan ntpd[224]: time reset -0.156714 s
Mar 22 23:08:57 dahan ntpd[224]: time reset -0.153315 s
Mar 23 06:24:57 dahan ntpd[224]: time reset -0.157639 s

So the alpha looks majorly bad, and the AMD isn't doing all that great
(however, I do have the alpha set up as one of its peers... perhaps
the alpha is pulling its ntp the wrong way?) Any ideas about what's
going on (especially with the alpha)?
-- 
Name: Dave Huang         |  Mammal, mammal / their names are called /
INet: khym@azeotrope.org |  they raise a paw / the bat, the cat /
FurryMUCK: Dahan         |  dolphin and dog / koala bear and hog -- TMBG
Dahan: Hani G Y+C 25 Y++ L+++ W- C++ T++ A+ E+ S++ V++ F- Q+++ P+ B+ PA+ PL++