Wake - Sleep Interaction -> Failed Recordings

[mikeh49]

My system is set to sleep after 5 minutes of inactivity.

On Thursday, 3-13, there were 3 programs scheduled to record at nominally 2200 with 2 minutes prepadding. I stopped watching a recording and deleted it at about 2152. It looks like the system did some housekeeping and went to sleep at 2157.38, and did not awake for the scheduled recordings. At 2152, NRecord had set a wakeup request for 2156 for the first program and 2158 for the other two.

On 3-14 when the system woke for EPG update, these recordings were marked as "Recording Service not running at recording time".

Event Viewer shows NRecord handled a power event at 2157.38.

It looks like there was a "race" between sleep and wake that was won by sleep.

Is there anything that should have prevented this, or could something be done? Can NPVR get the sleep time setting so sleep can be prevented if a recording is scheduled close to sleep time? Or, somesuch?

Logs attached.

[johnsonx42]

This is a known bug; if the sleep timer expires between the wake time and the recording start time, the system will sleep and the recordings will be missed. It's the sort of thing you'd think would happen so rarely that it'd not be worth the bother to deal with, but it seems to happen semi-routinely to some users. I have suggested adding a second "fail-safe" wakeup request at the recording time, so that if the system happens to sleep at the wrong moment there'll be a second wakeup but thus far there's nothing in NPVR to deal with this.

[mvallevand]

Yes I remember a lengthy discussion on this a few months ago http://forums.nextpvr.com/showthread.php...ording-why

Martin

[mikeh49]

Thanks guys. About what I thought. Also, Martin, I thought I remembered a previous discussion, but you're much better at tracking this down than I am.

A backup wake request would be worth a try IMO. I could live with missing a bit of the intro.

It has happened a couple of times for me in the past couple of years. Would not be a problem for cable shows that are repeated many times. Network shows are another story. WAF took a hit since one of the programs was Scandal, good thing it wasn't Nashville. Watched on-line via browser, so had to tolerate commercials.

[Graham]

Would the problem be avoided if NextPVR issued a wakeup request for the time that the recording is due to start (after adding the padding time)?

So show is in the EPG for 14.00 and padding is two minutes ... NextPVR asks Windows to wake at 13.58. The gap where the machine is awake but idle disappears.

We can increase the pre-padding if we are concerned that the machine wake-up process will take too long and not start recording before the start of the show.

[johnsonx42]

Would the problem be avoided if NextPVR issued a wakeup request for the time that the recording is due to start (after adding the padding time)?

So show is in the EPG for 14.00 and padding is two minutes ... NextPVR asks Windows to wake at 13.58. The gap where the machine is awake but idle disappears.

We can increase the pre-padding if we are concerned that the machine wake-up process will take too long and not start recording before the start of the show.

under normal circumstances you want the machine to wake up a minute or two before any recording is intended to start. if you attempt to start a recording exactly as the system is waking, it may not actually start because the tuner may not be ready. so even the second "fail-safe" wakeup request I've suggested at recording time would be no guarantee, but in most cases it would be better than just letting the machine sleep through the whole recording.

[Graham]

Would the problem be avoided if NextPVR issued a wakeup request for the time that the recording is due to start?

While looking for something else ... I found this http://www.nextpvr.com/nwiki/pmwiki.php?...nced#toc29

Sub had implemented the option before I mentioned it ... why am I not surprised!

[jcjefferies]

It may take about 20 seconds to a minute for the server to shut down during which time it is deaf to the wake up signal. It happened to me several times one week about a year ago when one recurring recording finished around 15 minutes before another was due to start. Because of the unfortunate combination of post padding, sleep time, pre padding, and 90 second pre recording wake up the second recording was missed. I did intend asking my son to write a PHP script to check if any "Pending Recording" was due to start within 10 minutes so I could prevent the server sleeping but it hardly seemed worth the effort as it is such a rare event.

Chris

[johnsonx42]

The really stupid aspect of this is that Windows itself doesn't handle this better. When the sleep timer expires and the system begins preparing to sleep, shouldn't Windows check to see if there's any wakeup requests in the next 5 minutes and abort the sleep if there are? I would understand if sleep were some new concept and they hadn't worked out all the fine details yet, but this has been going on for about 20 years!

[Handy.Man]

Win7 doesn't even respond to the keyboard when going to sleep !!! :eek: