Bug 15046 - Pausing track for over 2 hours results in "unable to find track" message
: Pausing track for over 2 hours results in "unable to find track" message
Status: NEW
Product: Logitech Media Server
Classification: Unclassified
Component: Streaming From SlimServer
: 7.5.0
: PC Windows XP
: P3 normal (vote)
: Investigating
Assigned To: Andy Grundman
: test_automation
Depends on:
Blocks:
  Show dependency treegraph
 
Reported: 2009-11-04 10:31 UTC by Mickey Gee
Modified: 2010-03-09 06:13 UTC (History)
5 users (show)

See Also:
Category: ---


Attachments
Squeezebox Server log (1.49 MB, text/plain)
2010-01-24 13:56 UTC, Mickey Gee
Details
First log file for incident March 5 2010 (200.03 KB, text/plain)
2010-03-07 21:20 UTC, Mickey Gee
Details
Second log file for incident March 5 2010 (170.81 KB, text/plain)
2010-03-07 21:21 UTC, Mickey Gee
Details

Note You need to log in before you can comment on or make changes to this bug.
Description Mickey Gee 2009-11-04 10:31:07 UTC
Spun off from bug 14230. Here's the relevant comments:

1) Running SbS 7.4.1 r28715. Track was paused for perhaps 14+ hours on local MP3
48K sample rate track. When restarted, Now Playing displayed a message about
"unable to find track" on top line of SB3, and second line displayed full path
of track that it couldn't find. SB3 could still play the track and then
stopped. Most likely played until buffer was empty. I think it did continue
playing the track where it left off from the pause and did not go back to the
beginning of track, but not 100% sure.

2) Occurs with 7.4.1 r28715 on SB3 with a 48KHz MP3 track. Absolutely reproducible at my house if track paused for 7 hours. Possibly seen if paused for 4 hours, but not sure.

3) Seen after 2 hours of pausing track.  7.4.1 r28715 on SB3 with a 48KHz
MP3 track.

I used remote to navigate My Music->Genres and able to find track and pressed
Play on remote. Track started up again OK, display looked normal.

Running Windows XP SP2.
Comment 1 Alan Young 2010-01-19 23:48:33 UTC
I presume that the filestore was local, and not over the net. Any SbS log messages?
Comment 2 Mickey Gee 2010-01-24 13:56:47 UTC
Created attachment 6458 [details]
Squeezebox Server log

It happened again on Jan 24. Log of occurrence enclosed.
Comment 3 Alan Young 2010-01-24 23:08:53 UTC
Mickey, here is the perl code that generates that error:

		if ( !-e $filepath ) {
			$log->error("File missing: $filepath");
			return {};
		}

As you can see, it is really very simple. It is a just a test for the existence of the file using a basic perl filetest primitive. Can you check you system logs to see if there is any clue there?
Comment 4 Mickey Gee 2010-03-07 21:16:30 UTC
Seen with Touch running TinySC on firmware r8622. Was playing continuously from Touch and 5 connected players -- 4 IP3K, 1 Radio. Seen only on one SB3 while playing MP3 track, and not other players. It's now happened to me on Squeezebox Server running on Windows XP, and now running on top of Linux. So probably not operating system related.

Of course, the other commonality is that they were Harry Potter Audio CD MP3 tracks ....

Don't think the logs have much info, since it happened overnight. Attaching them nevertheless.
Comment 5 Mickey Gee 2010-03-07 21:20:39 UTC
Created attachment 6605 [details]
First log file for incident March 5 2010
Comment 6 Mickey Gee 2010-03-07 21:21:13 UTC
Created attachment 6606 [details]
Second log file for incident March 5 2010
Comment 7 Alan Young 2010-03-07 23:34:08 UTC
Mickey, you filed comment #4 on 07 March but the overnight logfile attached (comment #6) covers the night from 04-05 March. When did the error occur?
Comment 8 Chris Owens 2010-03-08 09:09:02 UTC
Mickey notes he sees this primarily with MP3.

Osama, could you have support please keep an eye out for this?
Comment 9 Alan Young 2010-03-09 05:56:36 UTC
Ah, I think this is the clue:

Mar  5 09:35:34 slimserver.pl: [10-03-05 09:35:34.4261] Slim::Utils::Timers::__ANON__ (260) Error: Timer failed: Negative length at /usr/squeezecenter/Slim/Formats/MP3.pm line 193.
Comment 10 Andy Grundman 2010-03-09 06:13:52 UTC
There may be a bug in the MP3 seeking code.  Mickey, can you send me one of your Harry Potter files, assuming they are all encoded the same?