shutdown message hp-ux 11.0

[ernest tamura]

We were testing Sybase 12 on an hp 9000 k370 running up-ux 11.0 when the system became extremely slow. It was determined to shutdown Sybase and the restart the system. On shutdown we received the attached message. We have corrected the Sybase problem and have not received the shutdown message again. What does the message mean?

[Jeff Schussele]

Hi Ernest,

/etc/utmp keeps all login & logout info for users. What the msg was telling you was that it couldn't write the logout info for the users attached to those ptys.
Could indicate corruption in the /etc/utmp or not. Could have meant that it couldn't write to the file as well. Not a *big* problem unless it happens consistently, then I'd probably null out the file to rid it of any corruption.

Rgds,
Jeff

[Sridhar Bhaskarla]

Reboot will try to send a message to all the users on the system by using the corresponding pty devices. If there are corrupted entries, it will fail to write to them and displays them as errors.

I would not be worried.

-Sri

[Jeff Schussele]

Sorry - /var/adm/wtmp contains logins/logouts. /etc/utmp is just a user to pty map.
Again not a big problem unless it happens all the time.

Rgds,
Jeff

[Kent Ostby]

Generally speaking, this can be corrected by zeroing out the /etc/utmp file . But I would only do this if the problem reoccurs on subsequent reboots.

Best regards,

Kent M. Ostby

[Steven E. Protter]

I don't think the messages are associated with the performance issues. I suggest, as always data collection. See script attached.

SEP

[Naveej.K.A]

As all of them have mentioned, you really need not worry about those messages. This is caused due to corrupted /etc/utmp entry (utmp contains a record of all users logged into the system). The system on shutdown will send a broadcast message to all users currently logged in the system. This information is taken from /etc/utmp.
Hence this error message.

If you might have done a who -u just before your shutdown, you might have noticed these users logged in and killing these users with either kill -9 PID and fuser -k device file would have been impossible.

With best wishes
Naveej

[generic_1]

Hello I have seen these messages on my superdome lately. It seems to be delaying my reboot time noticeably. Although they aready reboot slowly due to the vast ammount of memory and processors. If you work in dark windows like me this is not good. Does anyone know what the timeout value is on these in 11i? Also what is the best way to fix this. Is a patch needed or is there a supported script from HP to add to the shutdown process to fix these? Thanks.

[generic_1]

Here is an example error.
reboot: Couldn't write to /etc/utmp entry: /dev/pts/11
reboot: Couldn't write to /etc/utmp entry: /dev/ttyp2
reboot: Couldn't write to /etc/utmp entry: /dev/pts/12
reboot: Couldn't write to /etc/utmp entry: /dev/pts/13
reboot: Couldn't write to /etc/utmp entry: /dev/ttyp3
reboot: Couldn't write to /etc/utmp entry: /dev/ttyp4
reboot: Couldn't write to /etc/utmp entry: /dev/pts/16
reboot: Couldn't write to /etc/utmp entry: /dev/pts/17
reboot: Couldn't write to /etc/utmp entry: /dev/pts/18
reboot: Couldn't write to /etc/utmp entry: /dev/pts/19
reboot: Couldn't write to /etc/utmp entry: /dev/ttypa

System shutdown time has arrived
reboot: redirecting error messages to /dev/console
Connection closed.