Re: Black Screen
- From: Bill "Wild Willy" Kredentser <wwilly@xxxxxxx>
- Date: Thu, 13 Oct 2005 13:13:18 -0400
> ** Reply to message from Mike Luther <mike.luther@xxxxxxxxxx> on Thu, 13 Oct 2005 15:24:15 +0000
Oops. Sorry my idea didn't help.
> But if you do that, the other objects in the STARTUP Folder don't load
> properly at that point. In other words, what I think I am seeing, is
> that an object which throws an error, even of what should be a
> non-related issue from the others in the STARTUP Folder, corrupts the
> operations relative to the subsequent work on what is in that folder!
Whoa! That's bad. Seems to me I've had broken objects in my Startup folder &
they didn't block the rest of the folder from running successfully. But it's
been a long time since that happened so maybe there's something that's been
introduced in the system via maintenance over the years that brought this on.
I don't have any broken objects in there now. And I don't think I want to
introduce any. <:-))
> Perhaps this is where your real issues with improper PEER operation come
> from, and we're seeing that your technique that does work there cleans
> up something in the environment which is sorta related to what I see
> here.
The thing with Peer is that the default configuration puts the NET START REQ
command into STARTUP.CMD. This is a bad idea. What I discovered, on both a
LAN-connected system as well as a dial-up, is that the NET START overlapped
with some rather CPU-intensive Desktop initialization during bootup. From time
to time, I would get a hard lockup requiring the reset button (Ctrl-Alt-Delete
wouldn't work). And this is with the NET START, before the LOGON had even been
issued. (Of course, on a dial-up, you don't want to be starting Peer before
your dial connection to the LAN has been established anyway, so you don't
necessarily even want to be executing NET START at all on such a box.)
Which brings me to my point. NET START REQ should be removed from STARTUP.CMD
& issued from a script invoked via a Program Reference object in the Startup
folder. Said script should issue a Call SysSleep of some duration before the
NET START, duration being determined empirically by watching how long it takes
for Desktop initialization to complete & the system to quiesce, something that
depends on your CPU speed (among other things). And on a dial-up system, you
shouldn't be executing NET START from the Startup folder at all. (Unless it's
part of some automation that ensures the dial out completes first.)
In all this, I make sure my .BAT that just exits runs at the "right" time from
the Startup folder. My Startup folder launches a handful of scripts which
start with SysSleeps of varying lengths so things get started in a particular
sequence, more or less one at a time. Desktop initialization is a delicate
time for OS/2 & you need to let it complete before you do anything meaningful
with the system. Multi-tasking after that is great, but it's not so great
during system startup. At least, that has been my observation.
I'm not sure if any of this is relevant to your situation, Mike, or the
original poster's.
--
WW
Team OS/2 Cincinnati & PROUD OF IT!
.
- Follow-Ups:
- Re: Black Screen
- From: William L. Hartzell
- Re: Black Screen
- References:
- Black Screen
- From: Kicklighter
- Re: Black Screen
- From: Wild Willy
- Re: Black Screen
- From: Mike Luther
- Re: Black Screen
- From: Wild Willy
- Re: Black Screen
- From: Mike Luther
- Black Screen
- Prev by Date: Re: Black Screen
- Next by Date: Re: Black Screen
- Previous by thread: Re: Black Screen
- Next by thread: Re: Black Screen
- Index(es):
Relevant Pages
|