NRecord using 100% CPU

[Astraeus]

Hi all,

I've got a bit of an odd problem which I'm trying to track down. At some point, but I've not been able to find a pattern, NRecord will start using 100% CPU and this will never end, so it kills the computer pretty much until I restart it (once it reaches this state, the computer is unusable).

I'm not really sure how to go about debugging this or trying to work out what's causing it, so I'd greatly appreciate any help.

The only clue I've been able to get is looking at the threads of NRecord, and the part that seems to be using 100% is:

hcw_mcmpeg2mux.ax!DLLCanUnloadNow (if the DLL can unload now, why hasn't it unloaded...)

I'm guessing this has something to do with the SoftPVR setup - is that a lost cause now?

Thanks a lot!

[sub]

What tuner and CPU does you machine have?

You'll probably need to post the logs next time it happens.

[Astraeus]

It's a Pentium 4 2.8GHz machine (so only one core, unfortunately), but before it goes wrong, it runs fine - the machine is dedicated to NPVR and isn't running anything else.

I've attached my NRecord.log from when it happened - it looks like it's logging at debug level, so hopefully something helpful in there? I couldn't spot anything obviously wrong in there though.

I'm using a Hauppauge USB PVR2 device

Thanks!

[sub]

These logs show it processing the overnight EPG update, which takes a few minutes from 2:01 - 2:07. There is no sign of it doing anything after that.

[johnsonx42]

I believe his problem was here:

Code:

2012-10-21 06:05:00.166    [DEBUG][7]    Stopping recording (9797 on 20). Past end time of recording. 21
2012-10-21 06:05:00.277    [INFO][7]    AnalogRecorder stream count now 0
2012-10-21 06:05:00.277    [DEBUG][7]    Graph stopping...
2012-10-22 02:01:42.225    [DEBUG][8]    Pre EPG Update
2012-10-22 02:03:12.253    [DEBUG][8]    About to update EPG

It hung trying to stop the analog graph. After that, NRecord did nothing but EPG updates.

[johnsonx42]

also, sub, I notice when reading his logs that he's got back-to-back recordings on the same channel with one analog tuner. It seems to allow post-padding, but not pre-padding in that scenario. Is that a bug in the logic? Shouldn't pre-padding be possible too?

[Astraeus]

I believe his problem was here:

Code:

2012-10-21 06:05:00.166    [DEBUG][7]    Stopping recording (9797 on 20). Past end time of recording. 21
2012-10-21 06:05:00.277    [INFO][7]    AnalogRecorder stream count now 0
2012-10-21 06:05:00.277    [DEBUG][7]    Graph stopping...
2012-10-22 02:01:42.225    [DEBUG][8]    Pre EPG Update
2012-10-22 02:03:12.253    [DEBUG][8]    About to update EPG

It hung trying to stop the analog graph. After that, NRecord did nothing but EPG updates.

Unfortunately, this does seem to be the problem.

The 100% CPU usage issue does not happen for a few minutes - once it triggers, it stays happening for days, even weeks, until the computer is rebooted. It's not the EPG. It's the Analogue jamming.

I'm not sure how to go about trying to fix this? I'd greatly appreciate any help

Thanks

[Astraeus]

Sorry to bump this post again, but I've done everything I can think of to fix this. I've ruled other factors other than the analogue recorder. Is this a hauppauge problem or an NPVR problem? Should I give up hope? (Digital is not an option for me, as we only have cable in this area, and cable is only available via a STB) :-(

[sub]

Is this a hauppauge problem or an NPVR problem?

I don't know. There is no obvious cause for the problem shown in your logs.

Do you have another set of logs you can post so we can see if the problem appears the same as the last log?

[Astraeus]

Sorry for the late reply - I'm only at home from time to time, so can only check things out then. I've come back home again, to again find NRecord using 100% CPU :-(

Again, it broke stopping the analogue recording, after which NPVR did nothing but EPG updates.

2012-11-30 21:35:00.750 [DEBUG][7] Graph stopping...

I've attahced the log

Thanks :-)