Subject: DEC 3000/400, NetBSD 1.4.1 fatal kernel trap
To: None <>
From: Ray Phillips <>
List: port-alpha
Date: 11/25/1999 15:59:02
Dear port-alpha:

My DEC 3000 - M400 machine which has NetBSD 1.4.1 on it and is really only
running Apache 1.3.9 and Squid 2.2STABLE5 crashed unexpectedly a few days
ago.  This is the information that was on the console after the last logout:

useswapctl: /dev/sd0b: Device busy
setting tty flags
starting network
default              JK-Gateway           done
Library-Mac          Webster-Multigate    done
jkmac34              Webster-Multigate    done
jkmac42              Webster-Multigate    done
jkmac67              Webster-Multigate    done
jkmac70              Webster-Multigate    done       Webster-Multigate    done
configuring network interfaces: le0.
add net default: gateway
adding interface aliases:
starting system logger
checking for core dump...
  syslogd: bind: Address already in use
savecore: no core dump
starting rpc daemons: portmap rpc.bootparamd.
mount_kernfs: : Operation not supported by device
starting nfs daemons: mountd nfsd.
setting securelevel: kern.securelevel: 0 -> 1
checking quotas: done.
building databases...
clearing /tmp
updating motd.
standard daemons: update cron.
starting network daemons: inetd.
starting local daemons:machdep.unaligned_print: 0 -> 0
Running: squid -sY  >> /usr/local/squid/squid.out 2>&1

fatal kernel trap:

    trap entry = 0x2 (memory management fault)
    a0         = 0xc0
    a1         = 0x1
    a2         = 0x0
    pc         = 0xfffffc000045c308
    ra         = 0xfffffc000045cb7c
    curproc    = 0xfffffc0004abe028
        pid = 29060, comm = newsyslog

panic: trap
Stopped in newsyslog at Debugger+0x4:   ret     zero,(ra)

Could anyone hazard a guess as to what caused the problem please?  I'm
rather keen to avoid a repeat of that incident.

Raymond Phillips