Folks I keep getting an error when exporting my timeline and wondered if anyone knew what this error is.
Computer Profile:
M2 Max MacBook Pro / 64 GB RAM / 2 TB / mac OS Ventura 13.3.1(a) / FCP 10.6.5
MacBook Pro (M2 Max, 2023)
Tangier, Davis' last post about disk subsystems got me thinking. Like you, I found there are almost no references to error -536870181. However -- converting it to hexadecimal = 0xE00002DB. Searching on variations of that hex number leads to the open source file IOReturn.h, which is part of the IOKit framework used in macOS. There we find:
#define kIOReturnNoSpace iokit_common_err(0x2db) // no space for data
IF that is correct, it implies some kind of out-of-space condition during an I/O operation. It's interesting the error happened when *exporting* a file (which obviously takes lots of space). That said, it might not be a normal "out of space" condition.
It could be some transient issue involving memory usage spilling to the page file (which takes more space but is quickly released). It could be APFS local snapshots consuming space, or even a problem with the drive itself.
Re the project repair function, I have a recent copy of your project (minus the media). I tried the repair and it found no problems.
We still haven't time-correlated the above errors from the system log with the "-536870181" error raised by FCP during export. We need to verify those are happening at the same time. Right now it's just a guess.
However I think the kIOReturnNoSpace error is a promising lead. An FCP user on Reddit reported that error -- also when exporting a file. He also reported logged errors referencing VTEncoderXPCService, which is the host process where the encoders run during export. That process itself can take up lots of memory, which can spill to disk in the page file.
Another complication is APFS has a "local snapshots" facility which can also consume space. I think it's supposed to release those when under pressure, but maybe the thread making the I/O call got tired of waiting and interpreted it as a failure.
I suggest you boot the machine in Recovery mode and run Disk Utility First Aid on each disk volume. To run it properly on the Macintosh HD volume, that should be expanded in the left sidebar of Disk Utility, then First Aid should be run from bottom to top, one pass *each* on "Data", "Macintosh HD", etc. Running it in recovery mode gives the best results because all other activity on the drive is stopped.
Disk Utility should update any space usage metadata that might be off. Then boot the machine, check disk space, take action as needed. Make sure there is *plenty* of extra space on all drives.
Once the system is up, I suggest running Disk Utility First Aid on all other drives. Observe the results of each pass, make sure there are no serious errors.
Yes, updated log only if you get the error again. I'd also like any FCP diagnostic reports from /Library/Logs/DiagnosticReports. You can go there in Finder via Go>Go to Folder (enter /Library/Logs/DiagnosticReports), then do CMD+F in Finder and search that folder for "Final Cut". Zip those files and send them to me, but only if you have another error.
If you have another error, please note the date and time (inc'l time zone) and let me know.
Tangier, that was reported by a Reddit user when exporting to ProRes from an M1 Max. He said it only happened if the source contained multiple lanes where both were visible on screen, e.g, PiP, etc. Versions not stated. Another user reported it when exporting to ProRes on an M1 Ultra on 10.6.2.
The Reddit user reported the console logs contained the below errors. You might check for those or similar.
One of them said it did not happen on export to H264. You might try that for comparison. If it is specific to ProRes output, that could imply some issue with the M2 Max ProRes engine or the software layers above that. It obviously should not happen.
The M1/M2 Max and M1 Ultra are the only Apple CPUs that have multiple ProRes engines. I believe they are used in parallel for cases involving multiple streams such as multicam. On ProRes export with sections of multiple layers showing, it's possible they are using one for decode and one for encode and there was some synchronization issue.
There was a previous "incorrect frame order" FCP bug introduced in 10.6.2, apparently as a side affect of an FCP performance optimization involving the M1 Max and Ultra CPUs. That was fixed on 10.6.4. Maybe this is a variation on that.
Console errors reported by Reddit user:
VTEncoderXPCService ERROR AppleProResHW (0xd3c24c88): GetEmptySpotInFrameInfoArray(): Timedout, could not find an empty spot
VTEncoderXPCService ERROR AppleProResHW (0xd3c24c88): ProResEncoder_EncodeFrame(): Could not get space in FrameInfoArray
Tangier, I just took a blind stab at reproducing it on my M1 Ultra running FCP 10.6.5 and Ventura 13.3.1(a), unfortunately without success. I tried three 4k/23.98 ProRes 422 multicam layers, each scaled to show output on the screen, total duration about 5 min. I exported that to 4k ProRes 422 and 1080p ProRes 422 both with and without pre-rendering the timeline. I never saw the error and nothing in the console log.
Unfortunately, the error does not indicate a frame number or position on the timeline where it happens.
Please check your console logs for errors, whether your timeline is showing multiple layers on screen, and how long of a timeline is required to reproduce it. If you can duplicate the project, then cut the timeline in half, then 1/4, etc. and still see the error, that would tell use more. If you can get it to that point, try removing all the effects and see if it still happens.
I worry about symptoms like this -- if it's a thread safety or synchronization error related to the multiple encoders, that is exactly the type of bug that can slip by testing.
Thanks Joe. I'll take a look at everything you mentioned and evaluate my timeline. I have never had problems exporting until yesterday. The first time it was a frame error. I quit FCP, deleted a few markers and it worked. Then I started getting this error. I need to think about the editorial changes made yesterday that may have brought this on. There weren't many at all. This export error comes with trying to export a ProRes Proxy. Timeline is in proxy preferred playback mode.
The timeline is about 98 minutes long, set at 3840x2160, 23.98 fps. It has a mix of 8K Red down to 6K and 4K Blackmagic video. This particular timeline is using proxies made from the Red and BRAW Toolbox-created clips at eighth decode quality. I tried exporting the first 30 minutes and that failed very quickly. I'm going to try exporting a few random short segments of the timeline and see if I get successful exports.
The only difference between being able to previously export my timeline and changes I made yesterday are:
I thought that perhaps it's due to my BRAW Toolbox clips (of which I have many) potentially causing the problem. I've had a lot of problems working with those clips that I have never seen before in all of my years of using FCP (X) am getting help from the developer Chris. I intend to post about this later perhaps on FCP.co to get some additional minds on some very strange anomalies happening to see if it's the M2 architecture, Ventura, or FCP, or something else.
Thanks Joe. Actually my first failure was using the Apple Devices 720p profile. Then I switched to ProRes Proxy with the intent of using Compressor afterward to make the streaming version. I will run another export shortly and see what the logs say.
I just tried a proxy export from a 60 min timeline containing a mix of 4k ProRes 422 and 8k Red R3D. It worked OK, so the problem scenario must be more involved. I will next add in some BRAW using the toolbox.
Tangier, have you tried making a custom preset in Compressor and making that available in FCP, and then using that preset to send to Compressor?
Tangier, I added stabilization to several clips, used BRAW Toolbox to add a bunch of BRAW clips to the R3D and ProRes stuff, set the FCP viewer to "proxy preferred", exported the 60 min timeline to 1080p ProRes 422, also tried Apple Devices 720p, and it still didn't fail.
That is on M1 Ultra. I guess I could try it on M1 Max, and maybe I could try it tomorrow on M2 Pro.
Ran some export tests today. There was one failure and I didn't grab the console information. For some reason things seem to be working right now and I can't think of anything that's changed. I should be happy, but I don't like that this can happen at any time and I don't know what triggers this behavior.
I wanted to test if the issue was writing to different destinations; one of my source drives or to the MBP's SSD.
Tests (exporting ProRes Proxy) in order:
Setup: MBP / Two USB-C 3.1 Gen 2 drives connected directly to MacBook Pro M2 Max vis Thunderbolt 3 cables
(later in the day)
Setup Change: Same MBP / Two USB-C 3.1 Gen 2 drives connected to MacBook Pro M2 Max via two USB-C cables through a CalDigit TS4 Dock, connected to MBP via Thunderbolt 4 cable. MBP in clamshell mode, monitor out to 34" LG Ultrawide monitor (mentioning all of this in case the monitor or CalDigit has an impact on this issue)
Due to success I haven't tried Compressor method yet.
Tangier, that CalDigit dock is powered. I bet that's the difference (well, guessing it is). What are the specs of your external drives? Seems like maybe the MBP isn't providing enough power or the TS4 just provides more... I heard of similar situations. Hmm.
Davis, I've been exporting on the setup with the CalDigit TS4 without issue until yesterday. The CalDigit is powered and the drives are Glyph Blackbox Pro desktop drives given to me with the project media; one 12TB and the other 18TB both with their own power.
Tangier,
Oops. Well, I've been having a few more "misses" than "hits" when trying to sort things, lately... ;-)
I've never heard of Glyph until you mentioned them (doesn't mean anything). Depending on how old those drives are, and how much they've been used, the cables could be an issue. I've heard reports of "bad" cables being shipped with various drives.
I avoid using anyone else's drives when working on a project. Too many unknowns. So, I almost always load up media on my RAID or SSDs, in order to be sure things will work properly. Years ago I was bitten a bit by the bad power supply brick issue that many LaCie drives had (and was fortunate to not have any catastrophic failures).
I have seen that specific error code mentioned for drive-related issues, but it isn't very specific or helpful.
Power cycling all gear sometimes helps to clear up some issues. Beyond that and standard troubleshooting, I've got nothing more to offer. I hope you get that sorted, sooner than later...
Cheers.
Tangier, I examined your console logs, and there are some SQLite errors. You can look at the timestamp and see if that correlates to one of your problems. First a few comments:
I loaded your "Anxious-TC Cut 3c" project (albeit without data). 1 hr 38 min, 1,764 clips. Big but not that big. Very nice looking timeline from an organizational standpoint. It behaves much faster than the previous versions I examined.
I examined your library and the Cut 3c project with a SQL database editor. I ran the SQLite commands PRAGMA integrity_check and PRAGMA foreign_key_check. They both returned no errors, but that is a low-level check.
Re your system log, the wording implies FCP sent a SQL query which exceded the capability of SQLite. Each question mark "?" is part of a parameterized query that in a later phrase of query construction will be replaced by a value. However SQLite must reserve memory for those, and there is a limit. It didn't run out of memory, it was probably rejected by the SQL parser for submitting to many parameters.
One possibility is there was some hugely complex timeline structure which led FCP or the CoreData framework to build and submit that query. My cursory look at your timeline didn't show anything obvious. Another possibility is FCP just made an error or became "deranged" due to some other issue. E.g, maybe it obtains the number of placeholder "?" it's supposed to send from some other object or data structure, and that was not initialized properly, causing it to send the excessive number of parameters and cause the error.
When it says "I/O error" I don't think that means a system-layer error. I think that is boilerplate verbiage reporting that the SQLite call returned FAIL.
Please send me an updated system log file (I only need that one file). You should be able to get that at /private/var/log/system.log. Zip that and send it to me.
Please make note of the time of day (inc'l time zone) of any major errors you get and what you were doing at the time, so I can compare to the timestamp in the log.
2023-05-08 20:53:59.481504-0700 0xd6cf3 Error 0x0 29715 0 Final Cut Pro: (libsqlite3.dylib) [com.apple.libsqlite3:logging] too many SQL variables in "SELECT 0, t0.Z_PK, t0.Z_OPT, t0.ZFLAGS, t0.ZIDENTIFIER, t0.ZNAME, t0.ZTYPE, t0.ZCATALOG, t0.ZMETADATA FROM ZCOLLECTION t0 WHERE t0.Z_PK IN (?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,
2023-05-08 20:53:59.482209-0700 0xd6cf3 Error 0x0 29715 14 Final Cut Pro: (CoreData) [com.apple.coredata:error] error: (1) I/O error for database at /Users/tangierc/Desktop/Anxious/01 FCP Projects/Anxious Build 3b.fcpbundle/_Anxious TC Edit/Anxious-TC Cut 3c/CurrentVersion.fcpevent. SQLite error code:1, 'too many SQL variables'
2023-05-08 20:53:59.514762-0700 0xd6cf3 Error 0x0 29715 14 Final Cut Pro: (CoreData) [com.apple.coredata:error] error: SQLCore dispatchRequest: exception handling request: <NSSQLFetchRequestContext: 0x600132b48000> , I/O error for database at /Users/tangierc/Desktop/Anxious/01 FCP Projects/Anxious Build 3b.fcpbundle/_Anxious TC Edit/Anxious-TC Cut 3c/CurrentVersion.fcpevent. SQLite error code:1, 'too many SQL variables' with userInfo of {
NSFilePath = "/Users/tangierc/Desktop/Anxious/01 FCP Projects/Anxious Build 3b.fcpbundle/_Anxious TC Edit/Anxious-TC Cut 3c/CurrentVersion.fcpevent";
NSSQLiteErrorDomain = 1;
Hey Joe, would it be possible for the "derangement" of FCP occur because of a HW issue?
How about the "hidden" Project Repair function in FCP?
The two people I've dealt with a couple/few years ago with the same error message had drive issues or bad cables. I realize that the error message isn't very specific.
Tangier and Joe, I wonder if having the Library on a separate drive from the media has any relevance here?
Thanks for the input Davis. I purchased all new cables with my MBP and the CalDigit. I too wondered if one of the drives was the issue, but they don't seem to be faulty. One of them has just ejected on occasion. However it's unclear if this error pertains to media specifically on one of the two drives I am using.
Power cycling was done as well.
All insight appreciated. Sometimes many of us such as myself who typically know to think of these things simply miss a beat here and forget to do the simple things, so all help, checks and reminders are welcome.
Fortunately we have Joe involved since he has a technological perspective deeper than I think the average person would be have about the workings of the tools we use.
Joe thanks for this detailed analysis. I couldn't find definitive information on this error code. I did wonder if this was a memory handling problem, but wouldn't know how to diagnose that.
Something I noticed is that when I try to export (share) my timeline, the share window loads relatively quick, but any successive attempts takes a significantly longer time for that share window to present itself. Very strange behavior doing the same operation. This was happening when I tested exporting the first 3,7,10, 60 and so on minutes of the timeline.
Assuming (but not certain) the timeline is processed linearly, the error would come up pretty quick in the export process; especially when I did the 10 minute. Though these were tests after the initial error which seemed like it didn't happen until a high percentage into the export process and therefor the timeline.
Just to be clear, do you want an updated log file only if I get the error again?
Final Cut Pro error -536870181
