shutdown command

[Louis McCurry]

Recently added the sept 01 patch set to a hpux 11 system. Since that time, when I issue the command: shutdown -r -y 0 the system goes to single user mode but does not reboot.
Before today the /etc/shutdown.allow file was blank. I added + root this morning but have not tried to shutdown and restart.
Any other ideas?

Thanks

[someone_4]

my shutdown.allow has always been blank.
what happends when you do
shutdown -ry 0
i know its the same thing but do you get the same result?

Richard

[Roger Baptiste]

A zerobyte (blank) /etc/shutdown.allow file or a missing shutdown.allow file should be
good enough to allow root
to shutdown the box. There
is no need to add any entry
in it, as long you want only
root to bring the box down.

Regarding the system coming
to single user state, what does the console logs show?
You can try running the same
command again from the single
user state.

Also, make sure the patches
were installed correctly!
Swlist -a and see whether
they are configured.

HTH
raj

[Darrell Allen]

Hi Louis,

Do you get any clues from /etc/shutdownlog, /etc/rc.log, or /etc/rc.log.old?

Does you system shutdown to single-user mode or does it try to reboot and comes up to single-user mode? Again, check the console.

Obviously and as Raj says, verify your patches are installed correctly.

Darrell

[Craig Rants]

Also, try /sbin/shutdown
Check to see if /etc/shutdown and /usr/sbin/shutdown are links and that no one has replaced them with shutdown scripts..

Just to cover all bases.

[S.K. Chan]

Are you in init level 3 ?
If you are, run ..
# init 2
# exit
Then login to the system again and shutdown.

[Louis McCurry]

Verified the pathes are installed correctly.

The system console had a text login banner (as opposed to the CDE).

I logged in and the system had 2 or 3 file systems mounted, /, /stand and /tmp (I think).

I issued the same shutdown command and a message similar to the following came up.
"disabling DHCP access...invalid, not configured" (This is not the exact message, but close. We do not use DHCP on this server)

It sat at this point for 7+ minutes, I eventually hit the power switch.

On power up the same DHCP message came up and stayed for 6 minutes, then the following messages came out:
"Synchronization of volume group vg0 complete"
"Synchronization of volume group vg1 complete"
The system booted normally.
The last lines in the rc.log.old file are:
Unload loadable modules
Output from "/sbin/rc0.d/K88kminit stop":
-----------------------------------------

The rc.log shows a normal system start.

There are links in /etc and /usr/sbin to /sbin/shutdown that is dated Apr 12 2000 and is 266240 bytes in size.

Thanks

[Alan Riggs]

Is it possible that someone has written a wrapper for the shutdown command in your environment? If that wrapper is not passing your command flags to /usr/sbin/shutdown that would explain the behavior.

Use "trype shutdown" to show where the shell is resolving your command. If it is anything other than /usr/sbin/shutdown then try using the full path to the command (and investigate the wrapper, of course).

[Bo Thide']

What if you use the 'reboot' command
instead?

Bo

[Louis McCurry]

No wrappers as type indicates the correct command is executing.

Is reboot functionally the same as shutdown -r -y 0 ?

Reading the man pages there seem to be no arguments and it immediately restartrs the system, which is fine for my needs.

[Uday_S_Ankolekar]

Hi,

Check /etc/rc.config.d/netconf file and look for dhcp, if anywhere you find DHCP_ENABLE=1 then make it to 0

Goodluck,
-USA..

[James R. Ferguson]

Hi Louis:

In answer to your question about 'reboot' vs. 'shutdown', 'shutdown' is an orderly transition which invokes the kill scripts at the various run-levels of '/sbin/rc?.d/'. As a last step in the process when a halt or reboot is requested (e.g. 'shutdown -ry 0'), then 'reboot' is called. Thus, invoking 'reboot' directly would bypass the shutdown scripts.

You indicate that your /etc/rc.log appears not to show any failues. The behavior you describe is consistent with something taking a very long time to terminate, or perhaps timeout.

Regards!

...JRF...

[Craig Rants]

Do you have Oracle or some other large application on this server? It can take a while for Oracle to shutdown the database if there are a lot of users in it at the time.

[Jej']

Hi,

Try to use reboot -h 0.
I think there's a link between your shutdown command and a shutdown script.

[harry d brown jr]

Shouldn't

/sbin/rc0.d/K88kminit stop

be

/sbin/rc0.d/K888kminit stop

???

could be executing out of order, or was it a typo ?

live free or die
harry