SUMMARY: serial console seems to change baud only in single user mode?

From: mark inaba <>
Date: Mon Feb 26 2007 - 11:17:42 EST
thanks to responders
Ric Anderson
Darren Dunham

good news: problem is gone
bad news: not sure why

i tried again to boot -sx, this time with some debugging echo cmds in
/sbin/rcS. however the machine came up to the root password prompt
fine. i did this with an /etc/rcS.d/S99foo file i made to sample the baud
rate of the console and save it to a file, and i also did it without it and
both times it came up to single user mode fine. i was curious to see
if the rcS script change might be the cause, or if it was just coincidental
to the problem vanishing...but i did not want to risk fs corruption by this
test. (and we have no other machines with this problem to test on).
perhaps it was a transient fluke, but the problem was persistent across
reboots of both the server and the terminal concentrator, and over 6hrs
showed consistent behavior. hmm..strange.



has anyone seen something like this or know where to fix it?

machine: v880
solaris 8
sun cluster  3.1
OPS (so udlms)
vxvm 3.5

when i try to boot this machine -sx it always produces nonascii after
"VxVM general startup".
I believe it has switched baud rate on the serial port because i consistently
get a y with two dots over it whenever i hit return.
i tried rebooting our terminal server, but the problem seems to be on
the v880 because
the behavior remains the same until the v880 is power cycled. then
everything is normal,
unless i do a boot -sx. after VxVM general start, it goes into
gibberish mode again.
what is interesting is that boot -x works fine, it boots all the way
up with no complaints...
however..when i do init comes down and then goes in the the
same gibberish mode
where returns produce y's with two dots over it. hoping someone might
have some insight into this behavior...

here's an example of boot -sx
ok boot disk works!!
boot -sx does not still
Rebooting with command: boot -sx
Boot device: /pci@8,600000/SUNW,qlc@2/fp@0,0/disk@w21000000871a4235,0:a
 File and args: -sx
SunOS Release 5.8 Version Generic_117350-05 64-bit
Copyright 1983-2003 Sun Microsystems, Inc.  All rights reserved.
Starting VxVM restore daemon...
VxVM starting in boot mode...
configuring IPv4 interfaces: ce2 ce3 ce4 ce5 ce6 ce7 ce8Feb 18
04:42:00 in.mpathd[57]: Failures cannot be detected on ce6 as no
IFF_NOFAILOVER address is available
 ce9 eri0.
Hostname: prdgc01a
VxVM starting special volumes ( swapvol rootvol var rootdisk_27vol )...
Not booting as part of a cluster
VxVM general startup...
Feb 18 04:42:41 in.mpathd[57]: Failure detection restored on ce6 as an
IFF_NOFAILOVER address is available

here's an example of multiuser going down to single user:

root@prdgc01a:/># init s

****    SYSCON CHANGED TO /dev/pts/1    ****
                                           root@prdgc01a:/># You have
running jobs

INIT: New run level: S
The system is coming down for administration.  Please wait.
remote filesystems: /vol nfs done.
                       Feb 18 05:21:09 prdgc01a /usr/sbin/vold[1692]:
problem unmounting /vol; Interrupted system call
Print services stopped.
                      Feb 18 05:21:10 prdgc01a syslogd: going down on signal
WARNING: ce0: fault detected external to device; service degraded

WARNING: ce0: xcvr addr:0x00 - link down
                        NOTICE: ce0: fault cleared external to
device; service available
        NOTICE: ce0: xcvr addr:0x00 - link up 1000 Mbps full duplex

Killing user processes: Agent: Terminating from signal 15
                                            WARNING: ce1: fault
detected external to device; service degraded
                             WARNING: ce1: xcvr addr:0x00 - link down

NOTICE: ce1: fault cleared external to device; service available
                                                     NOTICE: ce1:
xcvr addr:0x00 - link up 1000 Mbps full duplex
sunmanagers mailing list
Received on Mon Feb 26 11:18:17 2007

This archive was generated by hypermail 2.1.8 : Thu Mar 03 2016 - 06:44:04 EST