A kernel patch caused the machine to panic, leaving it in a confused state. I was able to boot from a CD and back out the patch, but it did not update the boot archive, even after issuing
/sbin/bootadm update-archive -R /a
to fix:
Boot into the disk failsafe, then run:
/a/boot/solaris/bin/create_ramdisk -R /a
ksh -o vi
export TERM=vt100
/usr/openwin/bin/resize
or, If booted from CD or in failsafe:
/a/usr/openwin/bin/resize
Problem: not able to connect to the console
-> start /CH/BL1/SP/cli
Are you sure you want to start /CH/BL1/SP/cli (y/n)? y
…just hangs indefinitely
Solution: We need to reset the serial port to the blade.
-> reset /CH/BL1/SP
Are you sure you want to reset /CH/BL1/SP (y/n)? y
Performing hard reset on /CH/BL1/SP
-> start cli
Are you sure you want to start /CH/BL1/SP/cli (y/n)? y
… connects successfully to the console now.
connecting to sun blades
start /CH/blade/SP/cli
start SP/console
for i in `ndd -get /dev/${DEV} \? | awk '{print $1}'| grep ^[a-z]`;
do
echo -n "$i:"; ndd -get /dev/${DEV} $i;
done
The problem is that the ssh session will wait for the command to complete and the ‘&’ to background the task does not seem to work. The net result is that the script running the nohup ssh command hangs.
This is because by default ssh does not allocate a TTY to run a command. You can force TTY allocation by using the ‘-t’ option to ssh, which will permit backgrounding, and the ssh command will return immediately.
In order to demonstrate, create a file /tmp/sleepy like so:
#!/bin/bash
while ( true ); do
echo `date` “sleep”
sleep 1
done
This command will invoke the script over ssh using nohup and &:
ssh -t localhost ‘nohup /tmp/sleepy &’
Tags: background, nohup output on remote, ssh
passwd -f <login>
XSCF> showdcl -v -a
DID LSB XSB Status No-Mem No-IO Float Cfg-policy
00 Running FRU
00 00-0 False False False
01 -
02 -
03 -
04 -
05 -
06 -
07 -
08 -
09 -
10 -
11 -
12 -
13 -
14 -
15 -
XSCF> console -d 00 -f
Connect to DomainID 0?[y|n] :y
#!/bin/sh
# ext-proto
#
# BIG BROTHER - PROTOTYPE EXTERNAL SCRIPT
# Sean MacGuire
# Version 1.9
# Mar 13th, 2002
#
# (c) Copyright Quest Software, Inc. 1997-2002 All rights reserved.
#
# Adapted to test RAID on Sun boxes (either raidctl or metastat or both)
# parmanand patram
# Nov 23rd 2010.
#
# SCRIPTS IN THE BBHOME/ext DIRECTORY ARE ONLY RUN IF
# THEY ARE DEFINED IN THE ENTRY FOR THE CURRENT HOST
# LISTED IN THE ext/bb-bbexttab FILE.
#
################################### SETUP #############################
#
# root must run a crontab like every 5 minutes to provide raidctl output
# 0,15,30,45 * * * * for VOLUME in `/usr/sbin/raidctl -l| grep Volume: | awk -F':' '{print $2}'`; do /usr/sbin/raidctl -l $VOLUME > /tmp/raidctl_${VOLUME}_txt; done >/dev/null 2>&1
#
# Additional setup is as 'normal' bb extension test.
#
# BBPROG SHOULD JUST CONTAIN THE NAME OF THIS FILE
# USEFUL WHEN YOU GET ENVIRONMENT DUMPS TO LOCATE
# THE OFFENDING SCRIPT...
#
BBTEST="RAID"
BBPROG=$0; export BBPROG
#
# BBHOME CAN BE SET MANUALLY WHEN TESTING.
# OTHERWISE IT SHOULD BE SET FROM THE BB ENVIRONMENT
#
# BBHOME=/home/sean/bb; export BBHOME # FOR TESTING
if test "$BBHOME" = ""
then
echo "BBHOME is not set... exiting"
exit 1
fi
if test ! "$BBTMP" # GET DEFINITIONS IF NEEDED
then
# echo "*** LOADING BBDEF ***"
. $BBHOME/etc/bbdef.sh # INCLUDE STANDARD DEFINITIONS
fi
#
# NOW COLLECT SOME DATA
# IN THIS CASE, IT'S THE STATUS OF THE RAID CHECKED BY raidctl AND metastat
# FOR MONTREAL, QUEBEC
# SELECT SOME LEVELS... GREEN IS THE DEFAULT...
WARN='undef' # this test will never have a warning. RAID is either OK or not OK.
PANIC='yes' # GO RED AND PAGE if FAILURE is set to 'yes'
# zero out the output file
cat /dev/null > $BBTMP/$BBTEST
# perform the raidctl tests
COLOR="green"
# IMPORTANT: user bb cannot run raidctl, so root must run a cron script to dump out
# raidctl info for each volume into /tmp/raidctl_VOLUME_txt
for VOLUME_FILE in `ls /tmp/raidctl_c*t*d*_txt`; do
VOLUME=`echo $VOLUME_FILE | awk -F'_' '{print $2}'`
TEST=`cat $VOLUME_FILE | egrep -v '^c|^Volume|Sub|Disk|-' | grep -v GOOD`
if [ "$TEST" = '' ]; then #disk is 'GOOD'
echo "OK for RAID volume $VOLUME" >> /$BBTMP/$BBTEST
cat $VOLUME_FILE >> /$BBTMP/$BBTEST
echo "-----------------------------------------" >> /$BBTMP/$BBTEST
else # disk has problems
COLOR="red"
echo "FAILURE on RAID volume $VOLUME" >> /$BBTMP/$BBTEST
cat $VOLUME_FILE >> /$BBTMP/$BBTEST
echo "-------------> ${TEST} <-----------------" >> /$BBTMP/$BBTEST
fi
done
# perform the metastat tests
for MIRROR in `/sbin/metastat | grep Mirror | awk -F':' '{print $1}'`; do
TEST=`/sbin/metastat $MIRROR | grep 'State:' | grep -v Okay`
if [ "$TEST" = '' ]; then #mirror id OK
echo "RAID mirror $MIRROR OK:"
/sbin/metastat $MIRROR >> /$BBTMP/$BBTEST
echo "-----------------------------------------" >> /$BBTMP/$BBTEST
else # disk has problems
COLOR="red"
echo "FAILURE in RAID mirror $MIRROR" >> /$BBTMP/$BBTEST
/sbin/metastat $MIRROR >> /$BBTMP/$BBTEST
echo "-------------> ${TEST} <-----------------" >> /$BBTMP/$BBTEST
fi
done
#
# AT THIS POINT WE HAVE OUR RESULTS. NOW WE HAVE TO SEND IT TO
# THE BBDISPLAY TO BE DISPLAYED...
#
# MACHINE NAME MUST EITHER BE A REAL MACHINE NAME, OR
# LOOK LIKE A REAL MACHINE (in the case of arbitrary measurements
# like temperature). IF THE NAME YOU ARE USING DOESN'T EXIST
# IN THE DNS THEN IT SHOULD BE LISTED IN THE bb-hosts FILE WITH noping,
# PREFERABLY IN IT'S OWN GROUP...
MACHINE="`hostname`"
#
# THE FIRST LINE IS STATUS INFORMATION... STRUCTURE IMPORANT!
# THE REST IS FREE-FORM - WHATEVER YOU'D LIKE TO SEND...
#
LINE="status $MACHINE.$TEST $COLOR `date`
RAID staus of disks
`cat /$BBTMP/$BBTEST`
# NOW USE THE BB COMMAND TO SEND THE DATA ACROSS
$BB $BBDISP "$LINE" # SEND IT TO BBDISPLAY
http://www.sun.com/software/whitepapers/wp-solarisinst/solaris_installation_deployment.pdf
and
http://www.sun.com/blueprints/1101/webstart.pdf
how to use Flash Archive from CD boot instead of from JumpStart Server:
http://www.sun.com/bigadmin/content/submitted/flash_archive.jsp
Basic boot commands:
http://www-uxsup.csx.cam.ac.uk/pub/doc/sun/solaris10/solaris10_basic_installation.pdf
