Locked out of project mid-work again

Resolve is now crashing to desktop 6-10 times a day, and every few re-launches, I also can't get back into the project (crashing/hanging at 95%+ and needing a full reboot dice roll), and then once every few projects it gets bad enough that I need support.

Changing timeline tabs is a regular culprit. But it can be anything from scrubbing a clip thumbnail to opening the inspector. It's like Resolve just collapses if the slightest thing goes wrong under the hood, idk how else to explain it. Error handling? Is that the term? I don't think I've gotten through a full workday without crashing in about a year. Losing several hours a week to this easily, often at crunch time.

This happens to be a cloud project, but this behavior is just as consistent on local ones. I then see that there were no backups of this one the whole time, despite this being launched from a template that had them enabled every 10 mins. I had a *library* backup from last night, only losing 3 hours of work, I can't restore it because I'm my library limit of 1.

I finally backed everything up and uninstalled deeply, did the full scrub down, reinstalled fresh, new database, try to open this project, loads to 100%, crashes to desktop first try. Every other project, working commercial stuff to utility sandbox projects, all fine. This is a short film I've been working in daily now. Very simple string outs and assemblies still, almost exclusively Ursa Cine 12K braw & H265 proxies.

Most recent logs below, and both a cloud & local export of the project with only about 5 mins of work difference between them. Any help appreciated.

https://www.dropbox.com/scl/fo/bwk8l2va ... o0kt9&dl=0

What video card/GPU?

What video card/GPU?

Hey Joe, see my signature, but also pretty sure(?) this is unrelated to video hardware based on everything.

Ah- on Tapatalk the signatures are hidden.

So the things you have that are somewhat different from most people are Sequoia and NFS. I’d think Sequoia is the more likely issue.

Is your sig accurate about running v19? Have you tried taking NFS out the situation? Perhaps making a secondary boot with an older OS and seeing if the crashes still reproduce in either of those situations?

Not saying these are easy but they might be informative.

This happens to be a cloud project, but this behavior is just as consistent on local ones. I then see that there were no backups of this one the whole time, despite this being launched from a template that had them enabled every 10 mins. I had a *library* backup from last night, only losing 3 hours of work, I can't restore it because I'm my library limit of 1.

Just to help eliminate the Cloud as a cause (and to get you by), what if you copied the project database to local and did it all local? That way, you could at least finish the project and ship it.

There may be things related to Cloud permissions or internet speed (or router issues), but those are complex enough that I'm not going to venture a guess.

This happens to be a cloud project, but this behavior is just as consistent on local ones. I then see that there were no backups of this one the whole time, despite this being launched from a template that had them enabled every 10 mins. I had a *library* backup from last night, only losing 3 hours of work, I can't restore it because I'm my library limit of 1.

Just to help eliminate the Cloud as a cause (and to get you by), what if you copied the project database to local and did it all local? That way, you could at least finish the project and ship it.

There may be things related to Cloud permissions or internet speed (or router issues), but those are complex enough that I'm not going to venture a guess.

Had the same thought, was able to import it briefly locally and made an export of that which is one of the files in that dropbox link. It’s since been borked as well.

Ah- on Tapatalk the signatures are hidden.

So the things you have that are somewhat different from most people are Sequoia and NFS. I’d think Sequoia is the more likely issue.

Is your sig accurate about running v19? Have you tried taking NFS out the situation? Perhaps making a secondary boot with an older OS and seeing if the crashes still reproduce in either of those situations?

Not saying these are easy but they might be informative.

Oh sorry, I never use that app so I didn’t know that! Good to know. Been on Sequoia since public release without issues across dozens of projects… this was going on before it as well. and ironically switching from SMB to NFS actually alleviated a handful of separate issues I had also been having and as much more stable now. (Just explaining why I ruled those out, not that you’re wrong!)

Interesting that switching the network made a difference. Are the problems that went away related? What’s on the network? Source files? More than that? What if everything is local to the machine?

One thing I’ve noticed about Resolve - it works well when nothing unexpected happens. But the error handling seems pretty abysmal. Like when you run out of disk space weird stuff happens rather than the app gracefully handling the error condition. So I’d expect a glitchy file service wouldn’t be handled gracefully either.

Interesting that switching the network made a difference. Are the problems that went away related? What’s on the network? Source files? More than that? What if everything is local to the machine?

One thing I’ve noticed about Resolve - it works well when nothing unexpected happens. But the error handling seems pretty abysmal. Like when you run out of disk space weird stuff happens rather than the app gracefully handling the error condition. So I’d expect a glitchy file service wouldn’t be handled gracefully either.

Short version is the NAS stuff is unrelated. Done too much troubleshooting and process of elimination across my setup over months of this. Resolve code is the problem, without question. SMB, NFS, internal storage, external drive, server, cloud, complex projects or simple, any footage, any length, doesn’t matter, it collapses to desktop at the drop of a hat, across any troubleshooting variables. As far as unexpected, agreed, and my system is virtually unchanged for like 3 years now. The issue has persisted among small changes (as in using different cache locations or .point updates or changing from USB to PoE for my panel). I should track it for like a week.

Update: my wife's MacBook Air has been the other collaborator on the Cloud project, so I just tried opening from that and it's fine. Same project that I was locked out of yesterday morning (both locally and in the Cloud).

I can't work on her computer though, just a troubleshooting note. I'll try exporting a new .drp from that though, as well as exporting all timelines out just in case. This *will* persist into future projects, so still in search of a proper fix.

UPDATE 2: Spoke too soon. After relinking all the media, I went to name the project to remove the CLOUD suffix as the final step before making a clean export, and it crashed when committing the rename and now I can't get back in again, even if I re-export from her side.

UPDATE 3: Okay, exported all the timelines one by one and imported them into a completely fresh project, once I started rebuilding the assembly timeline scene by scene again, I noticed simply pasting in S4 triggered the crash. So then I went back and copied each clip over shot by shot into the assembly timeline and it appears the culprit would be a few seconds of music that had a reverb plug-in on it. I turned off the reverb plug-in and it pasted it over fine. I was able to reproduce it both ways multiple times. It's what was blocking the whole project from loading at 100%, despite it being here for a couple days just fine. It's Waves RVerb. Used in my timeline and channel strip placement all the time no issue, this time it didn't like it. So yes, Resolve error handling. Wonder what it'll be next time! ha. Thanks for the help y'all!

Three years ago I made this feature request which, with the addition I included, would have solved your problem:
viewtopic.php?f=33&t=168458

Three years ago I made this feature request which, with the addition I included, would have solved your problem:
viewtopic.php?f=33&t=168458

Yesss. In my past aspirations as a musician and producer, 3rd party plug-in error handling was CRUCIAL in DAWs. I think the audio engine could just use an entire overhaul. (What is with the edit timeline essentially freezing any audio effects from equalizer to reverb, necessitating a toggle of something to refresh it, and not live update?!)

Anywho. Thanks Joe!