[tjwebster]

Hi there,

2 days ago my xbmc installation on my ATV2 suddenly stopped marking completed movies as watched and no longer resumes from the middle of an episode.

This did not happen immediately after upgrading to RC2 but instead a day or two later.

Any suggestions would be greatly appreciated.

Cheers,
Tristan

[ChipsAhoy]

This recently started happening to me too, but not all the time, and not from every device. I've been trying to gather enough info to file a report, but because it's so inconsistent, this has been a slow process.

I use a shared mySQL Database. It seems that it's ok after a fresh reboot, but starts failing over time. This could be why it has never happened from my iPad2, but seems to occur often on the ATV2. I can still mark episodes as watched manually, even after it starts failing.

Are you also using mYSQL or just a local database? Anyone else seeing this?

[tjwebster]

It is a local database and a hard reset of the ATV2 seems to have provided temporary respite. If anyone else is aware of a permanent fix then I would be glad to hear it.

[jmarshall]

The debug log will tell you what is going on as per usual.

[ChipsAhoy]

OK, thanks jmarshall, here's the info:

AppleTV2,1, Version 5.0.1 (Build 9A406a) wired ethernet
mySQL database running on Synology Diskstation 211j NAS
media is also stored on the Diskstation and shared via SMB
There are no playback issues.

XBMC 11.0 RC2 Git: 20120229-f38655f (compiled March 1 2012)
Installed and Updated via Nito,

Once it begins, the problem is consistent, but it sometimes doesn't occur immediately after a fresh reboot. After awhile (a few hours to a few days). watching a show doesn't automatically mark it as watched, and stopping a show in the middle doesn't save its place. No problem marking a show as watched or unwatched manually.

Note that I also watch media on the same network wirelessly from my iPad 2, which is also running RC2, and have never seen the problem there. I'm not positive the issue began with RC2, as I only used RC1 very briefly in this location and I may simply have not noticed it. It definitely wasn't there with the Eden betas or prior.

xbmc.log: http://pastebin.com/Sep7QA5N

[ChipsAhoy]

Does anyone who understands the logs better than I do see where the problem is? It looks to me like the log is saying that the watched count is incremented and committed proprerly, however the shows in question still are being displayed as unwatched.

[Ned Scott]

I think there's a way to get MySQL logs, but I'm not sure how. Will try to find something, unless someone beats me to it.

[ChipsAhoy]

Thanks Ned. I'll take a look at some of the mySQL desktop tools and see if there's a way to read the log. However, since tjwebster says he's having the same issue with the local db, I'm doubtful that will shed much light on the subject.

[jmarshall]

I think the issue is these:

17:25:43 T:127053824 ERROR: staticThread - thread Jobworker, Unhandled exception caught in thread process, attemping cleanup in OnExit

These occur due to invalid files perhaps (thumb extraction?) - you don't have debug logging on at that point - if you turn it on you'll be able to determine which files they are. Don't delete them ofcourse - we may need them to reproduce the problem.

Reason is that an exception will cause the jobworker to bugger off, but the jobmanager doesn't know about it, so you now have far less jobworkers in the pool, thus any low priority job (such as saving of filestate) are killed off. Will have a think about how this could be fixed.

Cheers,
Jonathan

[jmarshall]

Should be fixed in 28f79a5, ready for Eden Final. Thanks for the report!

Cheers,
Jonathan

[ChipsAhoy]

Thanks for the reply Jonathan.

Since it sometimes takes several days before the problem occurs, I had left debug logging off until the problem started showing it's ugly head, in an effort to keep the log small. It sounds like that won't work in this case, so I'll go ahead and do a fresh restart and then turn debug logging on immediately.

Let me know if I should proceed differently.

---

Oops, looks like we posted simultaneously, so I didn't see your next reply that you'd fixed it.

Thank you!

[ironlung999]

Mine stopped working after upgrading to frodo too. I changed the drive mapping from upnp to volume and it started to work. Why upnp stopped I dunno...

[Ned Scott]

UPnP file paths are dynamic and change, so XBMC can't keep track of what has been watched or not. It's the same reason UPnP can't be used with info scrapers and library features.

[ironlung999]

UPnP file paths are dynamic and change, so XBMC can't keep track of what has been watched or not. It's the same reason UPnP can't be used with info scrapers and library features.

Excellent many thanks.