I am reasonably new to Flame.
Our environment is Baselight and Flame. We are soon getting very fast auto-tiered shared storage and adding more suites (up to 10) so I wanted to check on options for collaborative workflows.
I have watched some of @ALan tutorials on shared stone + wire.
Does Flame always need to access media through stone + wire? What’s to stop us using NFS and having all data sitting on the shared storage which is also accessible by Baselight?
Essentially how should we be looking to optimise 10 Flame suites and 5 Baselight suites in the most collaborative way.
I’m also looking into BLG workflows.
this sounds perfect for the BLG Plugin, never had a chance like this actually using it, would be perfect.
As soon as you don’t cache imprted media, flame will read it as a soft-import from the given location. Having that exact path on the network, all other flames will use them. Rendering it for realtime playback will store it onto the local database though.
The big bottleneck here might be the server itself, even when being fast. Some BLG workflow eg. might be all 10 Flames reading the rushes live from the server. The BL machines will (might) do so, too, and as soon as the BLG file is used, all flames still are starting the root from the rushes. This could be heavy traffic.
Depending on the job you are doing, you can def. speed that up.
For example, if it’s one job needing all of your 15 ressources, I doubt they need it all together. So you can still cluster them into a main BL and Hero Flame, using the same files. While the other Baselights are feeding the hero BL in their own file system from the rushes and the Hero flame is feeding the flames with new files after the grade.
There are really a lot of possible ways in that case, as it depends on the size and timing of the job.
Thank you for the info Hilde.
What exactly does frame need to render for real-time playback, and what can it play directly off the storage (assuming the storage is fast enough)?
I guess we could also look at doing a centralised stone + wire on a separate section of the storage.
I should also add that it’s generally one or two suites per project, and no more than that.
I have some concerns that you will be having that large of an installation from the start with no prior Flame exipirience . Most recommendation and information you will get here will be centered around ancient data island workflow or just plain wrong.
First look at what network speeds you have, then storage. Just saying tiered isn’t enough. Then look at what competent technical skill set is available to you. Do you have a dedicated top engineer. What virtualization solutions are you comfortable with. How adavanced of Flame workflow do you really want? This is actually the key, and with out insult, I think you don’t know.
I could do a chat with you if you want.
Thanks for the info. I’m going to send you a DM.