My desktop at work is CentOS 3.8, using the motherboard's onboard Intel 865 (82865G) graphics controller. Most of the day today I've been having sporadic instances where the screen simply goes black for about 2 seconds. It has nothing to do with whether I'm using the mouse or keyboard (or not) and the delay before it comes back on is not correlated to whether I *continue* to use the mouse or keyboard.
This has happened a few times before. Sometimes the system is up for weeks at a time without showing any sign of this, sometimes it re-appears every few days. I've tried shutting down and restarting X11, but that doesn't get rid of it -- it seems only a full reboot will do so. And the longer I wait to reboot, the more often it happens.
It doesn't seem to be related to the clock or to ntp (I've seen ntp clock syncs trip the screen saver in X11 before), I've tried shutting down ntpd without any change in the behavior and the system clock is not drifting from other machines on the LAN either way. There's nothing in the log files indicating any activity that might be connected.
Anyone have a clue where I might begin looking? Having to reboot a linux machine with only 6 days of uptime galls me.
It might be your monitor, try using a new monitor.
HTH Oliver
Bart Schaefer wrote:
Oliver Schulze L. wrote:
It might be your monitor, try using a new monitor.
I don't have a better idea than the monitor atm (tho the graphics "card" is another candidate).
When it's occurring, try a) Cycling power to the monitor, including disconnecting from the pc if possible, just in case). b) Exchanging screens with a colleague to see whether that eases the pain.
Note: I don't assert that monitors are hotpluggable tho I've been doing it for years without noticing a problem. The instructions for my new Acer screen at work are confusing, in one para they say the screen's not hotpluggable, the next is says it is.
On 9/6/06, Bart Schaefer barton.schaefer@gmail.com wrote:
Just in case anybody was paying attention to this ...
Evidently the blanking has something to do with the machine being on a KVM switch with another computer (which happens to be a Mac Mini, but that doesn't seem to be relevant as it was powered off). I thought I had ruled out the KVM before I posted the first time -- among other things, why would rebooting the linux machine fix a failure in the KVM? -- but the problem has not recurred since plugging the monitor directly into the computer.
This isn't 100% conclusive because I've gone days or weeks without this happening in the past, but it's the first time I've gotten the problem to go away without rebooting.
On Thu, 2006-09-14 at 21:27 -0700, Bart Schaefer wrote:
I was. But had nothing to offer. That is common here.
I have mine on KVM, NP. But, IIRC, when I first joined there were some postings about some KVMs causing this type of problem? Search the archives. I *think* you'll find this is not unknown and not even that uncommon. I seem to recall that some brands were mentioned as being more likely to have this problem? Belkin??? Don't quote me though.
I guess I should admonish you to search the lists first? I know it's more work, but the list archives are sometimes less silent than this place. And frequently faster to give you an answer on "oddball" problems.
<snip sig stuff>
HTH -- Bill
On 9/15/06, William L. Maltby CentOS4Bill@triad.rr.com wrote:
I guess I should admonish you to search the lists first?
I did, but all the KVM problems I found were mouse trouble, not video.
On Fri, 2006-09-15 at 22:22 -0700, Bart Schaefer wrote:
Good man! Following on the track though, you said that if you hooked the monitor direct, NP? If so, there is a reason to consider that maybe the same KVM problem that caused mice to die could be responsible? If you can beg, borrow, ... a loaner to test with, maybe it confirms whether or not that is the problem.
Also consider this. Maybe one of the ports on your KVM is bad? Have you tried rotating the connections to see if the problem travels (assuming its not the port that accepts the monitor connection itself, but that's possible too).
<snip sig stuff>
HTH -- Bill
If its a Belkin KVM screwing in the mouse is a common feature they are cr*p...
William L. Maltby wrote:
-
Bart Schaefer -
John Summerfield -
Oliver Schulze L. -
Peter Farrow -
William L. Maltby