MP FCP stalls on Sharing of Apple Devices

[ColdCase]

ElCapitan 10.11.6, FCPx 10.3.4, nMP 8 core, dual D700s.

Over the past month I’ve been noticing that once in awhile the share to Apple devices (h.264) starts off fine but quickly stalls, little CPU activity, goes nowhere. Stopping the Background task often fails. Quitting and restarting FCPX leaves some unfinished partial files in the destination. If I leave it go long enough I will sometimes see q failed share notification with: “Error: this is null!”

Sharing a master file has no issue.

Today it seems to have turned into a hard fault.

So before I spin my wheels reinstalling software and running hardware diagnostics, anyone see this behavior before? Any opinion on the best, most efficient way for an amateur to determine the cause? A quick google search turned up nothing relevant. There are some console entries to interpret.

I’m thinking perhaps something in the h.264 transcode path is broken. Its been awhile since I reviewed the nMP architecture but could this be a Graphics Card issue?

But Handbrake has no issue converting the master file. I should try compressor, but its late.

 

[keysofanxiety]

I’m not sure, though if you were to contact Apple they’d suggest updating to High Sierra so you can update FCPX to the latest version.

I’d probably recommend the same but I appreciate there are good reasons you may stay on an older OS.

 

[fhturner]

A relatively quick, easy test is to create a new, temporary user sccount on the system (via System Preferences). Log out of your main account and into this fresh one, then try a similar export again, preferably with a new, test FCPX Library as well. This will at least let you know if there’s some corruption with your user Library or temp/pref files causing it. Let us know if this has an effect or not...

 

[ColdCase]

Thanks

AppleCare is still in effect as the machine is less that three years old. I have an iMac thats up to date, but have stopped updating the nMP because it just works. For one, its version of Compressor generates usable HTTP Live Streaming files where there is those generated by the newest compressor are not viewable. Just haven't had the time to figure that out.

When I send the project to compressor, it encodes just fine. This is a simple 2 minute segment, no cuts or effects.

I may clean out he dust, it due, and run a hardware diagnostic as well as try the new user approach.
[doublepost=1537799381,1537795781][/doublepost]As an update, I created a new user and when using a new library it seemed to work fine. The old library had the same issue.
[doublepost=1537799878][/doublepost]Also when I created a new library in the current FCPX, copied the video over, and share, it hangs.

I got some work to do

 

[fhturner]

As an update, I created a new user and when using a new library it seemed to work fine. The old library had the same issue.
--- Post Merged, 15 minutes ago --- Also when I created a new library in the current FCPX, copied the video over, and share, it hangs.

Hmmm...so the problem is specific to the Project/Timeline? Do other Projects do this or just this one? Also, if other Projects are doing it, are they duplicated from a common source Project maybe?

 

[ColdCase]

No its any timeline/project in the library, just have to share using a apple device preset. Doing a master file share is not a problem. .

Using the Original problematic FCP, not the test user:

I duplicate the original media and then drag it into the new test library, create a project, do the same king of edit (shorten and a transition) and share, works fine (using the same destination drive). Seems something goin on with the library. Although perhaps I should put the test library on the original drive to eliminate a drive issue. There is plenty of space all around.
[doublepost=1537801848][/doublepost]Running the test library on the same disk as the problem library works fine. Looking more like a library problem....

 

[fhturner]

Yeah, maybe something corrupted in the XML and other library files. You might try copying over the Event or Project from the bad library to a new one...

 

[ColdCase]

Yeah, it is a 2TB 2000 clip library with maybe 100 small projects in 20 or so events. I may just have to brute force through it.

Well copying individual event to a new library doesn't fix anything. Copying the media over and a project doesn't fix it either.

Update:

OK, going whacky here .... Things seem to be back to normal. I opened a troublesome project, selected all, copied and then pasted into a new project. Shared that to an apple device and it worked as expected. I'm dreading going through the library creating the duplicate projects so, for grins I went back an spot checked a couple other troublesome projects and they now share as expected..... perhaps there was cache or state somewhere not reset until a cut and paste project was shared.... I dunno... a couple beers right now sounds good

 

[ColdCase]

Well, spoke too soon, today I'm seeing the same behavior. The background task stalls, eventually returning the "null" error.

I moved the library off to a machine running 10.4.3 and it seems to behave as expected there.