TrueNas Scale with multiple nodes in different locations possible?

[iCa]

Hi!
I´m worting in a VFX and animation studio in Sweden. We currently have a cluster from Compuverde wich is awesome. But PureStorage boght them and support is going to end. Im looking at replacing our current cluster with a new one.
And my best case scenario would be if we could have nodes in all 3 of our offices, connected via VPN.
So say for example that we had 3 nodes with 50TB each in every location. Giving a total of 450 TB space. Would a VPNconnection over a 1gbit/s line be sufficient for keeping data in sync with about 15 people working on them?
Would this be possible to do with Scale?

 

[csj]

If you don't care about performance in the slightest, then in theory it will work.

Given your scenario I would much rather have 9 nodes in 1 central location and only have 2 of your sites coming across the VPN.

Your performance would be vastly superior to the scenario you describe and you'd get the same resilience.

 

[c77dk]

How far apart are the offices? You have to factor in the latency, and just the latency from Malmö to Stockholm would be a killer (somehow my guess is you're in those two cities :P)
I would go with a central setup, and some smaller in the remote locations and then use replication or Syncthing to keep them up to date

 

[morganL]

Hi!
I´m worting in a VFX and animation studio in Sweden. We currently have a cluster from Compuverde wich is awesome. But PureStorage boght them and support is going to end. Im looking at replacing our current cluster with a new one.
And my best case scenario would be if we could have nodes in all 3 of our offices, connected via VPN.
So say for example that we had 3 nodes with 50TB each in every location. Giving a total of 450 TB space. Would a VPNconnection over a 1gbit/s line be sufficient for keeping data in sync with about 15 people working on them?
Would this be possible to do with Scale?

1Gbe is a very serious limitation on the ability to synchronously sync data between sites... especial large video files.
You might be better using async replication.. e.g ZFS replication. However this would mean that 2 sites cannot simultaneously write to the same files or dataset.

 

[iCa]

Thx for all great replies!
Yeah we are talking about Gothenburg, Malmö and Karlskrona. Main office is in Karlskrona.
So maybe a black fibre connection between the sites would be best practice? We need to have artists in different locations working on the same files?
Say that i need 150TB of live space that everyone could access all the time.
What would be your advice on setup? Redundancy is ofc crucial as well as performance.

 

[morganL]

I think you are better picking a primary site.... replicating to a secondary
Aim for high speed networking for workers not at primary site.
Some static data or temp data can be at secondary site.