btw i now installed 2026.2.1 on a test machine which immediately started behaving bad in the same way as my cursed machine, uninstalling the service pack → no more issues on that one box
i dont know anymore .. macs are pain
i guess uninstalling a flame version does not uninstall/reset the other components like S+W and Mpd and stuff. … more to dive down.
HA! completely WIPED 2026.2.1 and everything ADSK from the machine, re-installed 2026.2.0 and BOOM no more error !!
i like fusion, but dont like how its implemented in resolve fusion standalone is great
Right, I’ve only ever used the standalone app. I’m pre-dating the Resolve integration. I think it still exists and works. But most people know it as a tab in Resolve.
as I said before… we are sticking with 2025.x for the foreseeable long term future. It is the pain I know, not a bucket of surprises with each release.
I’m going to look into this now as well. Working a huge project on 2026 and it’s unbearable how slow it gets. It’s unworkable. The autosaves are killing me.
Duplicate a sequnce. Wait.
Copy something in the library. Wait.
Make a new reel. Wait.
I hate to say it but for me this is Anniversary vibes all over again. I’m literally waiting more than working at this point. Fresh reboot, Fresh restart. It doesn’t seem to matter.
This really sux, sorry to hear about all the issues folks are having with 2026
We installed 2026 for a few days a while ago then immediately rolled back to 2025 as it was obvious that there was something really wrong with it. We mainly use Flame as a compositing tool these days but even then the load times and unpredictable behaviour, along with the post here on Logik, was enough to make us roll back
Funny thing is, we have Flame, Nuke Studio, Resolve and Baselight. Resolve gives us the least grief. The cache system is terrible and I’d only trust the Fusion tab for the bare basics (Fusion Studio is reliable) but overall it is the most reliable out of all of them when it is a few hundred bucks for the perpetual compared to the prices charged for subscription for the others. There are plenty of things that shit me off about Resolve but at least you know what they are and can work with them. If you’re charging thousands per year this stuff just needs to work. I know there was a push to place Flame as the central client attended hub for post production but man, if I had convinced a facility to go down that path then experience everything I’m reading on here…. sheesh!! Wouldn’t be a good look.
Just had probably my worst experience in Resolve in a while. Granted it had to do with it’s cloud storage, which is new and unrefined, not the decades old color page.
Project we’ve been on since last Spring, now has Cloud file set of 1.8TB shit the pants. Triggered by one of the editors running out of disk space. We tried a few different things, and the project degenerated. Somewhere along the way I couldn’t even import a local file to an urgent edit - as soon as I imported a local file it confused itself with its cloud sync status and claimed the just imported file was offline and not be found anywhere.
Fixed my urgent edit by taking various render chunks and combining them in Premiere to ship.
Then reset the cloud sync status, but was unable to convince Resolve to just re-learn where everything is.
In the end I had to re-import our last backup into a new project, keep cloud sync turned off so at least I can keep working on the master. Then set up a separate sub-project for the other editors and re-synced 200GB back to the cloud of the master file set, and then we’ll do smaller sub projects as we go along just to contain the blast zone if Resolve has another blow-out. The tools to control the cloud status and sync are absolutely horrid.
This is the trade-off for me: Blackmagic moves fast, chases the latest codecs, accumulates the most features. But most of their features are half-baked at best. It’s like sitting in the right seat while your teenagers learns to drive. The Flames and Nukes of the world move slow, but what they build tends to at least work.
Now, it seem the Flame team missed the mark with 2026. Major architecture change behind the scenes, which was much needed. But somehow they didn’t stress test it with large scale projects apparently. Seems like an odd miss, as this one would be an obvious test case to worry about. Especially when moving to a database and different file structure.
Question for those experiencing problems still - are those with 2026.2.1, which supposedly has a hotfix in it, or are they still with earlier versions of 2026?
Anyway, I think the introduction of Postgres database support in Flame2026 and some architectural innovations to the project are a major step forward. What Flame may need to do is to carry out a deeper and more comprehensive traditional architecture reform, and support Postgres database more deeply, just like PostgreSQL’s support for the Baselight project, which is more thorough.
Flame’s numerous nodes also need to be innovatively integrated, eliminating outdated or no longer necessary old nodes, and iterating on commonly used nodes. For example, in terms of color grading, with the MasterGrade node, some old color grading nodes in the Matchbox can be completely eliminated, and then the MasterGrade node can be optimized. At the same time, more convenient and practical modern comprehensive color grading nodes can be developed, which can be learned from Baselight and Davinci.
From Flame2025 to Flame2026, there have been some project architecture innovations over the years, and there have been some unforeseen bugs that I think are inevitable. For users who need stable production of Flame, it is best to have a dual version transition. Use Flame2025 to handle previous and significant projects, and use Flame2026 to initiate new projects, starting with small-scale projects. I believe Flame’s project architecture will become increasingly efficient. Looking forward to seeing significant changes in Flame2027.
I have the same question. Also, Mac or Linux?
I believe that is generally the plan. During NAB when 2026 came out, there was discussion that this was the first of several steps to move towards a DB architecture. The project settings was just the first of those steps which likely will take multiple releases.
And you’re right, some bugs are to be expected. There may have been a few more than was good, and that has created a bit of a speed bump or set back to take next steps.
A few, totally unrelated, likely irrelevant, and completely anecdotal thoughts
The grass is greener on the other side because it’s fertilized with bullshit.
With progress comes excitement. Both good and bad.
Most of us are fortunate to be alive and moderately healhy.
All of us are either tremendously brilliant, reasonably charming, devilishly attractive, funny as hell, or profoundly professional parallel parkers.
It’s gonna be okay.
Indeed.
Hotfix and I’ve got issues on both OS’s.
Bad news.
FWIW my issues with auto saves, etc all disappeared when I installed 2026.2.1. Mac OS.
FWIW - it’s easy to be frustrated about this situation, and easy to expect quick solutions.
The reality is that complex software changes like this can contain insidious bugs that are hard to identify and then fix. Doing this work is both tedious and fascinating brain teasers.
I spent half a day yesterday chasing why a website that I moved to a new server failed to work. Many ideas, no easy answers. LLMs drawing blanks too. In the end it was 5 lines of PHP code that was syntactically correct, passed every test, and had worked fine for years on PHP 7, but suddenly broke in PHP 8, without being related to any of the new language features.
Sometimes it gets even more extreme. Remember the bridge collapse in Baltimore in 2024 (which I’ve driven over numerous times in years past)? The root cause: a small plastic label on a signal wire that resulted in this particular wire not being installed correctly and eventually coming loose in an inopportune sequence of events. All the holes in the Swiss cheese aligned that night.
If you’re curious: NTSB Report. I love reading NTSB reports; so much can be learned from them about building systems and solving problems.
For all hard-to-solve problems, like the performance degradation of Flame 2026 with the new Postgres DB and the new file handling, all the critical holes in the cheese have to be identified.
Let’s partner with the ADSK devs in finding them, instead of just being mad at them, or dreaming about Resolve.
Without a doubt Postgres is a very positive step forward. Backing up would become a hell of a lot simpler as well and for those who have tech support, the addition of another Postgres database will be something they are familiar with and not as bespoke so more supportable.
What I’ve found is once it all starts going slow is to start a new workspace, copy just what’s needed to the new workspace and run things as lean as possible.
Rinse and repeat.
I’m going to strip the Linux box down to the studs and start again when I wrap this show. The Mac’s are new installs so there’s nothing to be done there…
