X-no-Archive:yes
I would like to get some thoughts on the following exchange. I doubt my own
objectivity in "turf wars", and want to make sure that I am not simply being
petty. Why might a remote admin need to be notified when a system is taken
down during non-production hours for hardware repair / replacement? I can
think of a few reasons for being aware of any restart just on principle,
including assurance that someone did not take down the system inadvertently,
or that a system abort was handled by more than an operator deciding to
perform a restart. Those would be aberrations from SOP that no one would
approve of, but when I see that the system was restarted, I am going to have
to wonder why it was restarted.

Normally, there are no developers on the development system on Sunday.
However, batch processes may ftp files over the weekend, for example. It is
far from certain that the people involved in shutting down the system would
have known to look for these processes to ensure that they did not kill
them, although our ftp jobs are generally robust enough to recover and
retry.

We are the Andover staff; the data center is elsewhere, but we share
responsibilities for various areas with the data center staff. The
boundaries are not always clear. In many cases this works well, in other
areas there is some uncertainty about exactly who is responsible for shared
responsibilities, and in some areas, the disagreements can get colorful. For
instance, we are responsible for all application and DBA issues. Each site
is supposed to handle security for their own users / passwords, although
some users are used in common, such as manager.sys (this does get colorful;
I am the only one here, while seven staffers there have access). We
configure the third-party scheduler; their staff monitors it 24 x 365.

Greg Stigers
Senior Consultant
http://www.cgiusa.com

-----Original Message-----
Greg,
I'm not against you being notified, but I don't understand why you need to
be notified, or why you even need to know for that matter. HP Predictive
Support provides notification for events that HP and the data center are
responsible for. I'm not in favor of creating extra procedural steps just
for the sake of someone's curiosity.  The Andover and the data center staffs
don't need to be burdened with more busy work.

Bob

-----Original Message-----
From: Stigers, Greg [And]
To: Bob
Subject: restart

Thank you for your reply.

Gotta love that HP Predictive Support! Should we have been notified? If it
is not policy to do so, could it become procedure on these servers? Thanks.

-----Original Message-----
From: Bob
To: EXT-CGI--Stigers, -Gregory; HelpDesk
Subject: restart

The development 3000 phoned home and with a problem that resulted in a
Memory SIMM being replaced. The change record was ########.

        -----Original Message-----
From: Stigers, Greg [And]
Sent: Thursday, September 16, 1999 3:54 PM
To: Helpdesk
Subject: restart

Why was the development HP 3000 restarted last Sunday morning, September
12th?