Hang after reboot !!

[Ivan Azuara]

Hi,

I have a strange event on my hpux 11.23 server. When i reboot it, after all the volume groups were actived whithout errors, i receive the next message:

stty: : Not a typewriter
stty: : Not a typewriter
stty: : Not a typewriter
/sbin/rc[14]: There is not enough memory available now.

, and 40 minutes after continue the startup process activating all the services until the system is ready.

When i check the /etc/rc.log file i can't find some strange message.

Some idea ??

Thank's in advance !!

[Patrick Wallek]

What about /var/adm/syslog/syslog.log? Anything interesting there?

[Dennis Handly]

>stty: : Not a typewriter
>/sbin/rc[14]: There is not enough memory available now.

Hmm, I don't see anything on line 14 in my /sbin/rc?

I do see stty above it:
# Set termio configuration for output device.
/sbin/stty clocal icanon echo opost onlcr ixon icrnl ignpar 2> /dev/null

But any errors should be suppressed.
Can you attach your /sbin/rc?

[Laurent Menase]

yiu can put a set -x in rc to see what command is executed

[Ivan Azuara]

Well, i suppose that there are a strange messages registred into the syslog.log file:

Feb 11 13:09:24 siinb55 syslogd: restart

Feb 11 13:09:24 siinb55 vmunix: Found adjacent data tr. Growing size. 0x33ac00

0 -> 0x73ac000.

Feb 11 13:09:24 siinb55 vmunix: Pinned PDK malloc pool: base: 0xe000000100c54000

size=118448K

Feb 11 13:09:24 siinb55 vmunix: Loaded ACPI revision 2.0 tables.

Feb 11 13:09:24 siinb55 vmunix: MMIO on this platform supports Write Coalescing.

Feb 11 13:09:24 siinb55 vmunix:

Feb 11 13:09:24 siinb55 vmunix: MFS is defined: base= 0xe000000100c54000 size=

2072 KB

Feb 11 13:09:24 siinb55 vmunix: Unpinned PDK malloc pool: base: 0xe0000001080000

00 size=393216K

Feb 11 13:09:24 siinb55 vmunix: NOTICE: cachefs_link(): File system was register

ed at index 5.

Feb 11 13:09:24 siinb55 vmunix: NOTICE: nfs3_link(): File system was registered

at index 8.

Feb 11 13:09:24 siinb55 vmunix:

Feb 11 13:09:24 siinb55 above message repeats 2 times

Feb 11 13:09:24 siinb55 vmunix: NOTICE: mod_fs_reg: Cannot retrieve configured l

oading phase from KRS for module: cifs. Setting to load at INIT

This is the content of my rc file.

#!/sbin/sh
# @(#)B11.23_LR

arg=$1
arg2=$2

PATH=/sbin
export PATH

#
# Set termio configuration for output device.
#
/sbin/stty clocal icanon echo opost onlcr ixon icrnl ignpar 2> /dev/null

umask 022

get_scripts() {
state=$1
mode=$2
dir=/sbin/rc${state}.d
# Fix for DSDe444244:
# Use of unspecified \ls\ in /sbin/rc can cause rc script failure
# ls has been replaced with /sbin/ls
/sbin/ls $dir 2>/dev/null |
while read name
do
case $name in
${mode}*)
path=$dir/$name
if [ "$mode" = "S" ]; then
desc=`$path start_msg`
elif [ "$mode" = "K" ]; then
desc=`$path stop_msg`
fi
echo $path $desc
esac
done
}

if [ -f /sbin/rc.utils ]; then
. /sbin/rc.utils
else
init_list() {
echo $1
}
add_list() {
eval $1
}
run_list() {
:
}
end_list() {
:
}
fi

#
# If /etc/rc.config contains default information (first boot),
# /sbin/auto_parms will invoke /sbin/set_parms to remedy the situation.
#
# For 10.0 release, the default HOSTNAME is unset or an empty string.
# Assume a timezone if /etc/TIMEZONE does not exist.
#

TZ=EST5EDT
if [ -f /etc/rc.config ]; then
. /etc/rc.config

if [ -x /sbin/auto_parms ]; then
/sbin/auto_parms
else
echo "\nWARNING: /sbin/auto_parms does not exist"
echo "DHCP invocation skipped."
fi
else
echo "\nWARNING: /etc/rc.config does not exist"
echo "System name not set, default $TZ assumed."
fi
export TZ

#
# Set runlevel information
#
set `getrunlvl` x
new=$2 # new run level
old=$4 # previous run level

#
# Check to see if we are run from /etc/inittab, or from the command line.
# If run from the command line, set the old run-level to the new run-level.
if [ $PPID != 1 ]; then
old=$new

# If the new run-level was specified on the command line, go to that state
# instead.

if [[ -n $arg2 ]]; then
new=$arg2
fi
fi

if [ "$new" = S ]; then
new=0
tosingle=1
else
tosingle=0
fi

BOOT=0
if [ "$old" = S ]; then
old=0
BOOT=1
fi

#
# Process scripts
#
found=0
if [ "$new" -gt "$old" ]; then
# new run level is higher than old, so run start scripts in
# all intermediate run levels and in new run level.
if [ $BOOT = 1 ]; then
init_list "HP-UX Start-up in progress"
else
init_list "Transition to run-level $new in progress"
fi
typeset -i lvl=$old
lvl=lvl+1
while [ $lvl -le "$new" ]; do
get_scripts $lvl S |
while read name descrip; do
if [ -s "$name" ]; then
add_list "$name start" "$descrip"
found=1
fi
done
lvl=lvl+1
done
elif [ "$new" -lt "$old" ]; then
# new run level is lower than old level, so run kill scripts
# in all intermediate levels and in new level.
if [ "$new" = 0 ]; then
init_list "System shutdown in progress"
else
init_list "Transition to run-level $new in progress"
fi
typeset -i lvl=$old
lvl=lvl-1
while [ $lvl -ge "$new" ]; do
get_scripts $lvl K |
while read name descrip; do
if [ -s "$name" ]; then
add_list "$name stop" "$descrip"
found=1
fi
done
lvl=lvl-1
done
# If we're ending up in state 0 or S, run the start scripts for
# that state.
if [ "$new" = 0 ]; then
get_scripts 0 S |
while read name descrip; do
if [ -s "$name" ]; then
add_list "$name start" "$descrip"
found=1
fi
done
fi
else
# old and new run levels are the same. Assume that execution
# is from the command line and run start scripts for the current
# run level
init_list "Starting subsystems for run-level $new"
get_scripts ${new} S |
while read name descrip; do
if [ -s "$name" ]; then
add_list "$name start" "$descrip"
found=1
fi
done
fi
if [ $found = 1 ]; then
if [ "$BOOT" = 1 ]; then
run_list boot
else
run_list
fi
end_list
fi

if [ "$new" = 0 ]; then
case $arg in
"shutdown")
exec /sbin/sh
;;
"reboot")
/sbin/reboot
;;
"off")
/sbin/reboot -h
;;
esac

#
# If transitioned to real state 0 (that is, not state S) via init, halt.
#

if [[ $PPID -eq 1 && "$tosingle" -ne 1 ]]; then
/sbin/reboot -h
fi
fi

#
# Output message to indicate completion
#
echo
if [ $BOOT = 1 ]; then
echo "The system is ready."
else
echo "Transition to run-level $new is complete."
fi

And is correct, the unique entry about stty into my rc file looks like yours.

Regards.

[Dennis Handly]

>i suppose that there are a strange messages registered into the syslog.log file:

What's also in /etc/rc.log?

>This is the content of my rc file.

(It would have been better if you attached /sbin/rc.)

>the unique entry about stty into my rc file looks like yours.

Hmm, it is on line 13, like mine.

. /etc/rc.config

You might want to check for junk/backup files in /etc/rc.config.d/.

[Ivan Azuara]

There are my configuration files.

Regards.

[Ivan Azuara]

Another file.

[Dennis Handly]

>There are my configuration files.

I don't want to see a specific file but the list of files.
In particular you should know whether you left a junk file there.

[Bill Hassell]

To add to Dennis's recommendation, use this command:

ls -lart /etc/rc.config.d

This will show the most recently added or changed files. The one directory in HP-UX can never have any junk or test or older versions of files. The reason is that *every* file is executed in that directory. If things were working OK in the past, the file(s) at the end of the list are suspect.

If you must keep older versions of config files, make a directory such as "old" and then move all the previous versions to the old directory.