[AG-TECH] KVMs for AG?

[Mike Weaver]

The Belkin manual mentions (in small print) in a couple of places not to
move the mouse or use the keyboard when switching channels and that the
mouse will be unresponsive for 1-2 seconds after a switch.  We try to be as
careful about this as possible!  It also states not to use the keyboard or
mouse on the active channel if the computer is booting as this can cause the
computer not to see the device.

That being said, I can reboot a working machine on channel 1, switch to
channel 2, wait exactly 5.2371 seconds, work on channel 2 for a while then
switch back to channel 1.  Wait another 5.2371 seconds and find that the
keyboard (or sometimes mouse) doesn't work.  I know that mouse
synchronization in particular is a very difficult problem for KVM switches,
but I do not see these kinds of problems on other brands ( we also use Rose
and Avocent/Cybex).

Mike

----- Original Message -----
From: "Don Morton" <morton at cs.umt.edu>
To: "Tom Coffin" <tcoffin at ncsa.uiuc.edu>
Cc: "Mike Weaver" <weaver at er.doe.gov>; <ag-tech at mcs.anl.gov>;
<bernholdtde at ornl.gov>
Sent: Wednesday, April 03, 2002 9:49 AM
Subject: Re: [AG-TECH] KVMs for AG?


> Tom Coffin wrote:
> >
> > I have noticed that with these switches they somtimes
> > freeze the mouse and keyboard if you are sending mouse
> > commands while switching.
> > I was recently putting our switch through some rugged testing
> > and figured that I really need to be patient or use a
> > slower process while switching. I now make sure that I am not
> > touching the mouse while switching. Let me know if the mouse
> > and keyboard still freeze while trying this movement.
> > Not a solution - but a workaround...
>
> yep, completely unscientific analysis and fix but, same
> situation here :)  i've found that it takes a few seconds
> for the electrons to switch between two machines and, once
> i actually do the keyboard switch, if i wait about
> 5.2371 seconds, then i can usually start moving the mouse
> without having things freeze up on me.  so, i'm being
> sarcastic but, it IS true that just waiting several
> seconds before moving the mouse typically reduces the
> chance of a freeze up.
> --
>    Don Morton                   http://MRoCCS.cs.umt.edu/~morton/
>    Department of Computer Science       The University of Montana
>    Missoula, MT 59812 | Voice (406) 243-4975 | Fax (406) 243-5139