[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

RE: KVM



Thursday, June 24, 2004 7:55 AM Scot L. Harris sagely noted:
> 
> There have been numerous reports that on certain systems 
> there appear to be some screen savers (people generally point 
> to the GL screen savers) that cause a system to hang.
> 
> I don't think anyone has nailed this one down but I have seen 
> enough posts to believe there is something that is broken 
> under certain conditions.  Most of the time if they chose one 
> simple screen saver or just blank the screen the problem 
> seems to go away.
> 
> I have seen similar things on some of my systems.  In one 
> case I had a heat problem with a particular motherboard.  I 
> have since replaced the motherboard and have not had any 
> problems.  But I have seen this on another box running the 
> GLmatrix screen saver.  I don't believe this to be a heat 
> problem.  It happens so infrequently on this box however that 
> I can not report changing the screen saver has resolved this issue.

I agree, something's missing in the symptom list. 

One thing that occurred to me this morning is a practice I follow when
using KVMs that I know many (possibly most) others don't. That is making
sure that all the systems connected to a particular KVM are all running
at the same resolution with the same refresh rates.

Over the years, I have found that it takes a high-end KVM to be able to
manage hosts with different resolutions and refresh rates. The collorary
to this is that nothing will scramble the limited brains of a cheap KVM
faster than different hosts running different resolutions and refresh
rates.

It seems to logically follow (at least to me it does) that if the video
switching can be messed up this way then the mouse switching could too.

This could also be the screensaver tie-in as it is possible on many
screensavers, and the GL screensavers in particular, to set them up to
change the resolution when they kick in and set it back when they exit.

As a side note, I did have a similar problem with one of my early FC1
boxes. It was my firewall at the time and I was making changes to
IPTables to set up a VPN to a client firewall. I screwed up an
incremental change and both didn't watch the log nor verify the
integrity of my IPTables settings after the change. As a result it was
left wide open for almost two hours. I noticed it when I went to make
the next change and found the machine acting squirrelly. Luckily the
customer FW I was building was still onsite so I unplugged my FW and
inserted theirs while I rebuilt mine. After the rebuild (and ever since,
as I'm still running that same build) there have been no more mouse
problems.

Now I realize that's a rather drastic troubleshooting approach. I did it
because I needed to be able to vouch for the integrity of the machine
being my primary FW. However, if anyone who is experiencing this has the
luxury of being able to rebuild, it would be interesting to see if that
solved their problem. :)

Eric Diamond
eDiamond Networking & Security
eric<at>ediamond[dot]net



[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]