End of recorded show crashing SageTV w/350

[Opus4]

BTW: I'm planning to change the default STV so that the deletion prompt at the end of a recording is optional, since it was reported that nothing bad happens when that dialog doesn't appear. There is no guarantee that this will get done before the next beta is released (whenever that happens), but it is on the to-do list.

- Andy

[1972]

Originally Posted by Opus4

previous

responses

... among others. So far, that hasn't solved it.

- Andy

Thanks for those two links, I didn't knew that the stop-before-the-end-of-the-show workaround already have been tried... A pity it did not work

[insomniac]

This would be much appreciated. I would love to see if others get similar results as I do when removing the delete prompt.

Besides the possible effect, it seems like it would be a good option to have anyway

I.

Originally Posted by Opus4

BTW: I'm planning to change the default STV so that the deletion prompt at the end of a recording is optional, since it was reported that nothing bad happens when that dialog doesn't appear. There is no guarantee that this will get done before the next beta is released (whenever that happens), but it is on the to-do list.

- Andy

[jphhughes]

Like rchady I tested GP-PVR on my system as well with no EOF problems. Your reference thread goes back to April of this year. Perhaps things have changed. The version I tested I believe was 92.11. Their has been a recent new release and when viewing the forum I did not see any recent references to 350 EOF problem. But I agree with rchady SageTV should make a policy statement regarding this problem.

[1972]

Originally Posted by 1972

Thanks for those two links, I didn't knew that the stop-before-the-end-of-the-show workaround already have been tried... A pity it did not work

But wait.... I reacted a bit to fast...

How comes this workaround won't work in the software????? If I do it manually than it does work, so if it is done automaticly what is the difference????

(I know several people tried to fix it in their STV, but the mechanism required to much CPU usage, if I remember it corrrectly. Is this what you mentioned Opus4? Or has Frey also tried this workaround in the Sage core software?)

[Opus4]

The current code stops playback before the end of the file is actually reached, but to go along with the above-mentioned option to disable the end of playback deletion question, a future beta version will also be able to stop playback even earlier than it does now. Maybe stopping even earlier will help... especially if you stop playback on the 350 before it begins.

Again, there is no guarantee that any of this will provide a solution. I'm told that the 350's driver does not correctly synchronize all operations regarding the OSD when an app makes calls into the driver for OSD functionality. Every solution is just trying to get around this problem, but it doesn't appear to be solvable w/o driver modifications.

If you think you are frustrated, my guess is that any programmer adding support for the OSD on the 350 has gotten a bit frustrated by now...

- Andy

[insomniac]

I definitely appreciate having those options open up. For me, EOF is very rare, and luckily recoverable in most instances. It will dissapear as soon as the delete prompt is gone. For others in unworkable situations, maybe they will be lucky enough to get rare freezes, etc..

If its one programmer that implemented the 350 OSD, would it not be time to have some fresh eyes look at the issue? This always helps me when Ive thought myself into a box...

Your mileage may vary...

I.

Originally Posted by Opus4

The current code stops playback before the end of the file is actually reached, but to go along with the above-mentioned option to disable the end of playback deletion question, a future beta version will also be able to stop playback even earlier than it does now. Maybe stopping even earlier will help... especially if you stop playback on the 350 before it begins.

Again, there is no guarantee that any of this will provide a solution. I'm told that the 350's driver does not correctly synchronize all operations regarding the OSD when an app makes calls into the driver for OSD functionality. Every solution is just trying to get around this problem, but it doesn't appear to be solvable w/o driver modifications.

If you think you are frustrated, my guess is that any programmer adding support for the OSD on the 350 has gotten a bit frustrated by now...

- Andy

[eobiont]

I came upon this thread because I have recently upgraded to Toast 7 for Macintosh.
When I try to take MPGs recorded in Sage via an Hauppauge 150, the file converts up until the very end when I get an error - the MPEG stream is incomplete. This suggests to me that the problem is in the way that the MPEG is closed, and not a problem with the playback. I don't know if this is an apples to apples comparison, but in my experience it appears that SageTV (3.0.7) is not producing a fully compliant MPEG2 video file with regards to how the file is closed -maybe? Couldn't that also be the problem with playback on the 350?

[Opus4]

To return to this old problem that drives 350 users crazy... The latest beta (v3.0.11) has 2 new features you can try:

1. Core change: "33. Play seam advance in the media player can now be set by the property videoframe/play_seam_advance. This might be useful in helping alleviate the PVR350 EOF crash bug."
   
   This can be used to force the recording to be ended earlier than it does now. It is the time, in milliseconds, prior to the end of the recording when SageTV will stop playback. I believe the default is 250, so you could try setting it to 1000 or 2000 to try stopping 1 or 2 seconds before the end of the file.
   
   
2. STV change: "30. Detailed Setup -> Customize now has an option to ask to delete a recording when playback ends for a) manual deletable shows only, b) all shows, or c) never"
   
   I think that one is obvious... but you can stop the EOF pop-up from asking if you would like to delete the recording.

If neither of these help, I have no idea what else could be done.

- Andy

[insomniac]

I was so excited to see this. Thx for adding these options. Now all I have to do is get enough nerve to reinstall the latest beta, or at lease go install somewhere else and rob the stv from it.

I.

Originally Posted by Opus4

To return to this old problem that drives 350 users crazy... The latest beta (v3.0.11) has 2 new features you can try:

1. Core change: "33. Play seam advance in the media player can now be set by the property videoframe/play_seam_advance. This might be useful in helping alleviate the PVR350 EOF crash bug."
   
   This can be used to force the recording to be ended earlier than it does now. It is the time, in milliseconds, prior to the end of the recording when SageTV will stop playback. I believe the default is 250, so you could try setting it to 1000 or 2000 to try stopping 1 or 2 seconds before the end of the file.
   
   
2. STV change: "30. Detailed Setup -> Customize now has an option to ask to delete a recording when playback ends for a) manual deletable shows only, b) all shows, or c) never"
   
   I think that one is obvious... but you can stop the EOF pop-up from asking if you would like to delete the recording.

If neither of these help, I have no idea what else could be done.

- Andy

[rapidskies]

Originally Posted by insomniac

I was so excited to see this. Thx for adding these options. Now all I have to do is get enough nerve to reinstall the latest beta, or at lease go install somewhere else and rob the stv from it.

I.

I'd be very happy to see if either of these settings work as well. I've been living with this PVR350 EOF pain in my side for close to 2 years now and it is very frustrating trying to catch every program before it hits the end. My EOF bug happens about 90% of the time if I let the file reach the end. Anyway let us other suckers that bought the PVR350/SageTV bundle know how it goes.

[insomniac]

ever move the 350 to a different slot? Its quite common to have better or worse results from changing slots with certain troubled cards. If 90% failure is what u have now...you can do practicly anything to make your system better

I.

[rchady]

Just curious, has anyone been able to confirm whether or not the changes in the beta help the PVR350 problems? I am too chicken to try the beta unless I have some confirmation it has helped someone at least...

-Robert

[jrgarrett]

I just installed V4 and noted that I still had the EOF bug. Changing Setup/Customize/Ask to delete recordings at the end of playback to "Always" seems to resolve this issue on my system. Might be worth giving it a try

[insomniac]

ok...so i believe I may have found a magic bullet

let me start off by saying that I did not have any EOF issues until v3. After v3, EOF rared its ugly head with a vengeance. After being annoyed enough, I decided to play around with why it was happening. At times, when EOF happened, I was able to rewind or sometimes put sage to sleep (if caught within 1-2 seconds). I could even ffw to the end of a file and I would get the delete prompt immediatly(no EOF)..."hmmm...."

To me it seems like a buffer/IO type issue. So I tried a dogpile of settings and found some that worked for me! It sure would be nice if someone else gave them a try to see if they get most excellent results like me..

In sage.properties or client.properties file, change the setting: pvr350_demux_numbuffers to 16 - (not sure if this one is necessary, but it would probably be wise to do all the same settings that I did, and subtract one at a time if you feel the need)

in registry:

HKEY LOCAL MACHINE\SYSTEM\CurrentControlSet\Control\Session Manager\Memory Management

change DisablePagingExecutive from 0 to 1

make a dword value called IOPageLockLimit and set it to 40000 or so (which is conservative)

reboot, try it out, and post results!

for me, I believe the setting that worked was the DisablePagingExecutive setting. It makes it so that does not page data from RAM to disk. It is suggested that you have a minimum of 256mb memory. Secondly, ive heard that the iopagelocklimit setting was disabled in sp1. Although everytime i make this setting, i notice performance increase. Whatever...

There is 1 more setting I havent tried that may help, if the others havent:

LargeSystemCache=1

I suggest that you try the other settings before you do this. I have seen a few posts that this may cause some issues with some ATI graphics cards...

Your mileage may vary...

I.

[aberson]

I attempted just the "DisablePagingExecutive=1" fix (none of the others).

The result of this is that when I hit the EOF, instead of Sage locking up, now the computer reboots immediately.

With the EOF lockups, I had configured a button on remote to kill sage (and a .bat file to automatically restart it when it got killed), so I could hit that button to restart sage if I accidentally hit the EOF. This gave me an annoying downtime, but better than nothing, and I think better than rebooting.

Every so often, Sage would lock up in such a way that the process can't be killed and a reboot is necessary.

Bottom line is that changing this setting does seem to have an effect, just not a positive one for me.

[Steve2112]

I have found that during the EOF glitch, it seems that Sage is locked up, but it really is not. My work around has been to press control-alt-del, and bring up the task manager. That immediately lets Sage free... with the taskman open, I then delete the file. Then I close taskman until it happens again.

[insomniac]

If you could try the others for informative purposes, that would be great. Try changing the unmbuffer size setting (to 16) and see what results you get. If this does not work, it would be nice if you tried all of the settings that seemed to work so well for me. Maybe the pagingexecutive setting alone crashes you, but used with all other settings might end up being a good thing

Just for giggles, im trying out the largesystemcache setting to see if i get better or worse results.

I.

Originally Posted by aberson

I attempted just the "DisablePagingExecutive=1" fix (none of the others).

The result of this is that when I hit the EOF, instead of Sage locking up, now the computer reboots immediately.

With the EOF lockups, I had configured a button on remote to kill sage (and a .bat file to automatically restart it when it got killed), so I could hit that button to restart sage if I accidentally hit the EOF. This gave me an annoying downtime, but better than nothing, and I think better than rebooting.

Every so often, Sage would lock up in such a way that the process can't be killed and a reboot is necessary.

Bottom line is that changing this setting does seem to have an effect, just not a positive one for me.

[insomniac]

Largesystemcache seemed to have no improvement in avoiding the 350 EOF, but:

after playing a bit...i decided to change my numbuffer setting(in the properties file) back to 32. I went from very infrequent EOF 5 second freezes (in my book, thats fixed. or good enough!) to very frequent EOF hangs(20 seconds or more and required standard EOF covert maneuvers). I then changed this setting to 8. As of yet, I have not had any EOF hangs, but have only tested on 5 or 6 files.

Couldnt tell you what this setting does or what the minimum allowed setting is, but it definitely seems to make a difference either in conjunction to the other tweaks or even w/o. Although I had not tested this setting alone.

Jeff or someone with Sage,

Can you elaborate on this setting? From doing some forum hunting way back when, i recall jeff telling some users with a 350 to change the setting from what was default 16 to 32, which is now default.

Does changing this setting to 8 (as I did) even register? Seems to have, but could be placebo I suppose..

I.

[aberson]

Originally Posted by insomniac

after playing a bit...i decided to change my numbuffer setting(in the properties file) back to 32. I went from very infrequent EOF 5 second freezes (in my book, thats fixed. or good enough!) to very frequent EOF hangs(20 seconds or more and required standard EOF covert maneuvers). I then changed this setting to 8. As of yet, I have not had any EOF hangs, but have only tested on 5 or 6 files.

Is this setting only on sage 3? I don't seem to have it in my un-upgraded version.

I have another suggestion for a workaround for the EOF problem. It seems that I do not get EOF's if I have 2 consecutive recordings. I imagine this is because the file isn't "closed" until the end of the 2nd recording, so the 1st one doesn't EOF. [Do other people notice this?] My workaround suggestion is to always have something recording after a program ends.

My idea for how to implement this is to modify guide data such that there is a new episode of some fake show on every 30 minutes boundary (00:00, 00:30...). The show could be 1 minute long, and could be on some channel that I don't care about. I would then set a low priority favorite for this show, and tell it to only keep 1 episode. Bottom line is that it would always record this show immediately after every recording, possibly alleviating EOF's.

So, short of making this behaviour in an STV, is there a way for me to hack a new channel into the guide data??