John,
nv6000_smit trys a "soft" kill of the ovw and ovw_binary's that are
running.
It will wait up to 2 minutes to see if they have closed. Note: you will
see the "closing" message on your consoles. After two minutes it then
does
a "hard" kill if the soft kill did not bring it down. It appears that a
soft kill works more than 95% of the time, and is no different than a user
closing their windows.
I have been using this method with customers for over 2 years without
causing
other problems. I also require a database cleanup/sync process be run
every 2 weeks for large NetViews, every month for all other NetViews. So
if
this did cause a small database problem once in a while, at least for me it
is a non-issue with regular database maintenance.
Note: for those people running with NetView clients, a KILLEUIS does not
close the EUIs that are running on the clients. To get around this, I
copied
nv6000_smit to the clients (don't tell support!) and then wrote a script
that
runs at the server, figures out which clients are connected, then runs
nv6000_smit KILLEUIS at each of the clients before running it at the
server.
Steve Hochstetler
Tivoli Professional Services
steve
Date: Thu, 11 Jun 1998 16:58:56 -0600
From: "John A. Dorsey" <dorsey@COLQUITT.ORG>
Subject: Re: BACK UP
MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Steve and others,
-> It is very easy to "correctly" bring down ALL NetView users on a system.
-> Run the command
-> /usr/OV/bin/nv6000_smit KILLEUIS. It will close all NetView sessions.
-> After you do the /usr/OV/bin/ovstop,
->
-> [...]
->
-> Steve Hochstetler
-> Tivoli Professional Services
nv6000_smit KILLEUIS stops user sessions by just
killing the ovw processes. I have been under the impression
that, for read-write maps at least, this is not a completely clean
way of stopping EUIs.
I've been further under the impression (although I don't
have much to go on here) that a cleaner way of shutting down
is to forceably log off netview users through the security
GUI. (With security enabled, obviously.) Can anyone at IBM/Tivoli
comment on these superstitions?
John
|