Flame Project Server - Local

Is anyone using the Flame Project Server locally? How has that been for you?

What do you mean by “locally”?

As in opposed to being in the cloud. That was the primary reason for introducing it but I am wondering how well it is working when installed on a locally hosted server.

Uh…sorry… I have no idea what you are talking about. Define Flame Project Server?


There is an installed for a project server in the new release. I just assumed that there would be nothing stopping you from installing one locally. Perhaps I am mistaken?! I thought I had even read through something which talked about how to install it on a headless system but I possibly misread.

Based on what @ALan has demonstrated in the past and what we know about the pains of media duplication and the benefits of single point of ingress, egress (and regress via archiving), not to mention the security benefits of snapshotting the whole server instance whether remote or local, it seems like it’s a no brainer.

When we move to 2023 it’s top of my list to deploy.

Subscribe to my channel.


@cnoellert You don’t need to wait to upgrade you actual workstations to 2023. You can set up a Project Server even if your actual Flame version is lower. S+W is backwards compatible.


For sure, we’re just planning on switching it all over at once. Thanks man

Thanks Alan. I will definitely take a look.

dope man , i have a question on data-flow, so you have the tiny s+w server , and a nas, and no stones on the workstations, but you still have the stones as partitions/folders on the NAS and the workstations just mount that?

So i guess what I am asking is - If I work fully uncached can this work without any stone setup? Only leaving the project metadata on the sw server but have every flame just directly access the files from the NAS like resolve would do it?

You still need “stonefs” to render to and keep other caches and stuff, even if you are not Caching actual source plates. The stonefs’ can be on any storage you want, including your NAS.

We have a dedicated StoneFS NAS that is ~180TB of all SSD. Then we have an everything else NAS that is about ~700TB of 120x 6TB spinning HD.

But yes, you can reference material directly from a NAS as Soft-Import.

Workstation have only OS drive, and no actual data on them. No S+W defined on workstation, only Project Server. Think of them as “dumb computer nodes”. They are ephemeral. It makes system administration and flexibility way easier.

1 Like

yea exactly, so i would just setup a folder on my nas set it up as the framestore location in the s+w server and the flames use that as their stones directly? thats it?

in my case its just 2 flames mostly just 1 active at a time I just want the failsave aspect and snapshots for my projects. I would also build a fast ssd “stone nas” with nvmes.

yup, basically. you don’t define any S+W on the Flame itself.

I highly recommend you virtualize the S+W servers (not the NAS though) and do the “stonefs per project” workflow that we do. Its a bit more overhead, but makes things very safe and fast.

1 Like

I do a similar approach with resolve. Server is on a Proxmox container. Easy to back up restore migrate. And in theory remotely accesible through Tailscale. Lately have been experimenting with BM Cloud. Does Adsk have a similar service?

I do the same in resolve ( well I used a mac mini as project server but same thing) then I use pgdump to dump the whole library every 3 hours to googledrive.

But yea - proxmox + small container running s+w like alan sounds like a great plan.

own partition per project makes a lot of sense I tend to do that even for a local flame install as otherwise you have no idea whats eating data - just like how you do it with avid mediaComposer, one avidMediaFiles folder per project.

So theoretically , I could use a DAS solution on flame1 , and have that shared via smb like its a nas , and then flame 1 gets full DAS performance and flame2 uses that also as its stone via smb… ?

I would not do that, as then Flame1 becomes your point of failure. And you will be rebooting a Flame more often then your would a dedicated NAS. Get 25gigE networking, and basically that covers nearly all formats.

yea just wondering If i could set this up like this just to test the workflow.

what happens when the stone get wiped and all media is soft-imported - is there any other metadata in there at all? As in … do I need to snapshot/backup my stone folders/partitions or is it just meh and I just loose my timeline caches and all generated media?

Wiped intentionally or catastrophically?

Do you have a current NAS? Just use that for testing. The key really is the networking. All our NASes are backboned at 100gigs, Workstations at 25.

1 Like

does that matter ?

I just want to cover my bases and understand what I need to keep snapshots and nighlty backups and stuff of to avoid stupid downtime due to hardware failure.

I do have NAS, its all 10G right now, so no not enough as caches for the “hero flame”.

25G is definetely in-scope of things I want to do for sure.

Qnap TS-h1290FX or something along those lines seems nice as a highspeed framestore, right now its all 10Gbit/s synology but synology does not have anything with nvme storage.

this looks nice and then i just get those sonnet dual 25Gbit/s and with smb multichannel I should get pretty much thunderbolt3 line-speed or something like that… sounds dope tbh