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

Re: [Condor-users] Condor 7.4.4 hibernating with logged in local user



Ian,

I thought you were saying that the keyboard activity detection was not working. That's why I mentioned condor_kbdd.

However, going back to your previous message, I get the impression that keyboard/mouse activity detection is likely working fine. The problem is when a user is logged in but does not touch the keyboard/mouse for longer than the threshold that you have configured. The system then hibernates, and the user has to press the power button to bring it back to life.

One solution, which I mentioned previously would be to write a startd cron script that publishes the number of logged in users. Your ShouldHibernate expression could then require that there be no logged in users.

Another solution, as you suggest would be to hibernate in a state that does not require the user to press the power button. Unfortunately, I don't have any experience with the details of getting that to work, but it does sound worth pursuing.

--Dan

On 11/23/11 10:53 AM, Ian Cottam wrote:
Thanks to Dan (and others) for replies. I've also found Dan's earlier post
on this.
In that, Dan asked what we wished to do when a user was logged in. I did
at the time think - obviously, not hibernate. But, the reason for that is
the only combination of states we have found that works
(Condor/BIOS/Linux/. . .) puts the PC into such a deep sleep that the only
way for the local logged in user to get it back up is to press the power
button, which then causes a reboot (clearly not acceptable).

We are using the terminology "RAM" to Condor and "S1" to the PC's BIOS.
What other combinations could I try (especially if you have experience
with it working)? I have tried "RAM" and "S3", which did not work.

Thanks all.
-Ian
ps: not using condor_kbdd daemon, but to be honest, I can't see how that
would help.


On 23/11/2011 15:52, "Dan Bradley"<dan@xxxxxxxxxxxx>  wrote:

Ian,

Are you using condor_kbdd?  This daemon is used to detect X keyboard
activity.

--Dan

On 11/23/11 4:21 AM, Ian Cottam wrote:
The ShouldHibernate line below does not seem to be enough to stop our
Linux/64 PCs sleeping whilst in use by someone (logged in) locally using
the keyboard/mouse/screen.

HIBERNATE_CHECK_INTERVAL = 30
TimeToWait = 600
ShouldHibernate = KeyboardIdle>   $(StartIdleTime)&&   $(CPUidle)&&
$(StateTimer)>   $(TimeToWait)&&   (State == "Unclaimed" || State ==
"Owner")
HibernateState = "RAM"
HIBERNATE = ifThenElse( $(ShouldHibernate), $(HibernateState), "NONE" )

Any suggestions of what to add to prevent such?
Thanks
-Ian

ps: my apologies if someone already replied to this, but I can't find
my previous email on the subject
_______________________________________________
Condor-users mailing list
To unsubscribe, send a message to condor-users-request@xxxxxxxxxxx with a
subject: Unsubscribe
You can also unsubscribe by visiting
https://lists.cs.wisc.edu/mailman/listinfo/condor-users

The archives can be found at:
https://lists.cs.wisc.edu/archive/condor-users/