Duplicate Recordings

[johnsonx42]

please post your recording-backup.xml and recurring-backup.xml

[johnsonx42]

also, 1.9gb isn't much free space; that's less than 30 minutes of HD, maybe 2 hours of SD. you've probably been bouncing off the disk space limit for days as you record and delete recordings.

your database is probably hopelessly screwed up.

i always advocate having recordings on a separate physical drive; if that isn't possible, then they should be on a separate partition. letting the C: partition ever get full or even really close to full is very bad.

[nia]

OK - sanity checked the logs. The above apparently went a little too fast.

My recording disk ran full today at 21:28 (just before the last set of dancers in a dance show - not good for WAF :o ). That generated the error Johnson describe in the attached zipped "NRecord.log.1" twice in the logfile. But that is the only occurances I can find, and only today. I have checked earlier zipped logfiles - no occurrance of the disk full/database error.

Hence, this can IMO not in any way have impacted the scheduling perfomed over the past weeks resulting in doublets.

So - I have cleaned up the disk, rebooted, redone the "Update EPG", and the log is once again attached.

[nia]

please post your recording-backup.xml and recurring-backup.xml

Attached.

I know about having more disk space free - just been a busy couple of weeks week and it had gotten away from me

[johnsonx42]

OK - sanity checked the logs. The above apparently went a little too fast.

My recording disk ran full today at 21:28 (just before the last set of dancers in a dance show - not good for WAF :o ). That generated the error Johnson describe in the attached zipped "NRecord.log.1" twice in the logfile. But that is the only occurances I can find, and only today. I have checked earlier zipped logfiles - no occurrance of the disk full/database error.

Hence, this can IMO not in any way have impacted the scheduling perfomed over the past weeks resulting in doublets.

Sorry to be argumentative, but your past log files are chock-full of the SQL disk space error. It may or may not have anything to do with the double recordings, but I don't trust your database any more and I doubt sub does either (and honestly, neither should you!).

edit: ok, maybe "chock-full" is an over-statement, but it's a lot more than twice, and still doesn't change my opinion of your database's logical integrity:

Code:

C:\nia log temp>find *.* "disk is full" /c

---------- NPVR.LOG: 0

---------- NPVR.LOG.1: 13

---------- NPVR.LOG.2: 0

---------- NPVR.LOG.3: 0

---------- NRECORD.LOG: 0

---------- NRECORD.LOG.1: 2

---------- NRECORD.LOG.2: 0

---------- NRECORD.LOG.3: 0

[johnsonx42]

your recurring-recording.xml looks ok, so you should have no problem starting over with an empty database; then you can safely import the two .xml's and get your past and future recordings back. the only thing that won't be preserved is single (non-recurring) pending recordings, so you'll need to reschedule those yourself.

[johnsonx42]

your logs (both older and new) show database errors, probably due to missing and mismatched records:

Code:

2010-09-17 22:36:22.971    [ERROR][5]    Unexpected error in UpdateRecurringRecording(): System.NullReferenceException: Object reference not set to an instance of an object.
   at NShared.EPGManager.LoadListingsMatchingTitle(Channel channel, String title)
   at NShared.RecordingServiceProxy.UpdateRecurringRecording(RecurringRecording recurringRecording)

as I said, your database is toast.

[nia]

Posted a post that became obsolete due to Johnsons continued analysis

[nia]

Sorry to be argumentative, but your past log files are chock-full of the SQL disk space error. It may or may not have anything to do with the double recordings, but I don't trust your database any more and I doubt sub does either (and honestly, neither should you!).

I'm sorry - i certainly don't want to waste anyones time. I appreciate your efforts. And I will discard the database.

BUT

The "disk is full" only occurs in the files in the time window of less than 15 minutes tonight while the disk was full. The 13/2 disk full you refer to is just the system repeatedly trying to work until I freed up space. But this was from the space incident tonight.

So the preceeding weeks issues with double recordings can't be related to something happening tonight?!

What I see at EPG update in the morning at 7AM is several lines of

Code:

2010-09-15 07:02:57.217    [ERROR][8]    Unexpected error in UpdateRecurringRecording(): System.NullReferenceException: Object reference not set to an instance of an object.

This may be because the database is shot from earlier events. I have had some system freeze-ups apparently related to a faulty RAM. That may have introduced a problem leading to :

as I said, your database is toast.

But maybe it's a trace of a problem. I'll let you guys evaluate that.

In any case, I will do as suggested and recreate the database - I'll search for a howto actually do that, but if anyone has an authoritative link don't hesitate to post :-)

Anyway - I'm looking hard for a SQL database integrity analysis / repair tools to look if it is actually ill afterall and how ill (post mortem). Suggestions welcome.

[johnsonx42]

you can either just delete the database and copy in the new blank database provided "empty.db3", or run "NextPVR.exe -emptydb". the latter is probably simpler, since you don't have to shutdown and restart the recording service.