Ok,
Here it is the part that matters to you:
1) If the system cannot be interrupted due to this error state, forcibly
reboot it by cycling power on the CPU or executing a RS from the
CM> or SP> prompt. Watch the boot messages and interrupt the
autoboot sequence to perform an attended boot.
2) Boot from the primary boot path and interact with the
ISL.
On earlier 700's, this requires:
a) pressing the "escape" key to interrupt autoboot
b) type "boot pri isl"
On later 700's and all 800's, interrupt the autoboot by pressing any
key when the 10 second pause appears after the system displays the
CONSOLE path, and PRIMARY and ALTERNATE boot paths.
3) At the ISL> prompt type:
ISL> hpux -is
This will boot the system up in single-user mode. A shell prompt
should eventually appear.
4) (Optional) Type "/etc/bcheckrc" (linked to /sbin) to activate
volume groups and check all file systems.
If you are left with a "inbcheckrc" prompt, determine which file
systems need to be manually fsck'd and do so.
a) # mount -a <--- this reports if/which file systems need fsck
b) # fsck /dev/vgXX/lvolYY <-- this is the usual fsck command
Once completed, all file systems should be mounted.
5) Mount /usr, /var and /tmp to access commands and 'vi'.
6) Compare the last successful reboot date with files modifed
afterwards:
# tail -1 /etc/shutdownlog # check the date/time
Check to see if files were modified after that date:
# ll -tr /etc/rc.config.d | tail -10
# ll -tr /sbin/init.d | tail -10
# ll /etc/inittab
Example:
# tail -1 /etc/shutdownlog
12:49 Sat Jan 3, 2002. Reboot: (by eon!root)
# ll -tr /etc/rc.config.d | tail -10
--(snip)--
-r-xr--r-- 1 bin bin 731 Nov 13 16:00 cmcluster
-r--r--r-- 1 bin bin 4408 Jan 4 14:08 netconf
In this example, the netconf file is a candidate for error
Review file changes and correct as needed. Generic versions are
located in /usr/newconfig/ under etc and rc.config.d. The
/sbin/ commands have no duplicates.
Each and every day is a good day to learn.
