[mythtv-users] Duplicate entry for key PRIMARY in recordedmarkup

[Michael T. Dean]

On 03/04/2011 05:00 PM, Bill Meek wrote:
> On Fri, 2011-03-04 at 11:02 -0500, Michael T. Dean wrote:
>> On 03/04/2011 08:32 AM, Khanh Tran wrote:
>>> I upgraded to .24-fixes a few weeks ago and I'm starting to see the
>>> following database errors after some recordings.
> ...
>>> Duplicate entry '2030-2011-03-04 08:00:00-30-28215' for key
>> 'PRIMARY'
> ...
>> The only ways I could see this happening is if you have duplicate
>> channel IDs, which is only possible if your channel table is broken,
>> or if for some reason there are actually 2 recorders writing the same
>>   file (which shouldn't be possible--and the code to prevent such
>>   hasn't changed in a /very/ long time.)
> Khanh Tran, thanks for bringing this up and Mike for the detail.
>
> I have 136 of these in my logs from 8-25-2010 [trunk at 25815] to today
> [v0.25pre-1317-g624c23c].
>
> All of them are on my Hauppauge HVR-1600's analog tuner which is
> connected to my STB (Motorola DCT-700.)  Channel IDs 2xxx.
>
> The digital side of the HVR-1600 and the 2 halves of an HVR-2250
> are used for over the air recording. Channel IDs 1xxx.
>
> Most errors appear in pairs like:
>
>   Duplicate entry '2117-2011-01-18 22:01:00-30-46214' for key 'PRIMARY'
>   Duplicate entry '2117-2011-01-18 22:01:00-31-46214' for key 'PRIMARY'
>
> and all but 2 are type 30&  31 (MARK_VIDEO_WIDTH&  MARK_VIDEO_HEIGHT.)
> The 2 odd ones are type 11 (MARK_ASPECT_4_3.)
>
> I only looked at the most recent 3 cases where the recording would still
> be on disk, but is isn't. That's because the recordings had no audio and
> I deleted them and asked for MythTV to rerecord.
>
> I'd never tied these errors to my missing audio recordings.
>
> Does this look like a valid MythTV case, I'd be happy to collect more
> data and open a ticket.

Ah, I think I see what's happening.  These are all probably interrupted 
recordings (broken into multiple pieces), right?

I'll fix it up "soon."

(And, on the bright side, the particular confusion that causes this type 
of issue will be banished, shortly, with a major schema change.)

Mike