crash at 1:30am almost every night

[johnsonx42]

yes, no doubt I typed too loud again.

[johnsonx42]

it's baaaack! the 1:30am crash that is...

hopefully you've not been playing around with NPVRTSMon.ax since releasing 2.2.6. attached is the crash.mdmp, nrecord.log, and the 5-second zula patrol recording. I realize the crashed zula patrol recording is probably no use since it's just a regular recording, but it was so tiny it was easy enough to upload.

in the log you can see that it begins to try to end the zula patrol recording at 2am:

Code:

2011-09-01 02:00:00.692    [DEBUG][7]    Stopping recording (100940). Past end time of recording.
2011-09-01 02:00:00.692    [INFO][7]    DigitalRecorder.StopStream() handle: 6D

but then it dies, since that part of NRecord had actually already crashed 5 seconds after the zula patrol recording started.

[sub]

This dump shows it crashing while parsing some ATSC EIT entries. Its unknown what data was going by at the time the crash though.

I've attached a patch, that along with a HKEY_LOCAL_MACHINE\Software\NPVR\LogEIT=1, will create c:\temp\eit.log files to store the EIT packets it's about to parse. Hopefully with one of these I'll be able test the exact scenario here.

This cause the eit.log to be overwritten a stack of times per second, so hopefully it doesn't impact too much much on your performance.

[johnsonx42]

thanks sub, I will install that tonight before the overnight recordings begin.

[johnsonx42]

HKEY_LOCAL_MACHINE\Software\NPVR\LogEIT=1

I presume I was supposed to create that as a DWORD value?

edit: nevermind, I did it as a DWORD and I see it's creating the eit.log

[johnsonx42]

I got the crash last night, but I don't think the eit.log is going to be any help. It crashed at the beginning of the 1:30am Zula Patrol recording like always, but the other tuner was also recording so the eit.log continued writing until 2am. I've attached the logs anyway, just in case there's any new info in either one.

Can you make a version that writes the eit.log per tuner?

[sub]

Here is a version which should produce separate logs for each capture graph run. c:\temp\eit-xxx.log.

[johnsonx42]

great - that should do the trick. thanks.

[johnsonx42]

oh, this creates thousands of these files per minute. I though you meant it would create these once per recording graph, not one file for every individual group of packets. I'll try to let it run like this, but the drive may fill before I get any useful data.

[sub]

oh, this creates thousands of these files per minute. I though you meant it would create these once per recording graph, not one file for every individual group of packets.

That's what I thought as well - I'll check it in the morning.