DL6 multi branch setup - and usage of FQDN for nodes

Deadline
1

The performance improvements in Deadline 6 make it very easy to monitor remote farms with a local monitor which is great!

One issue comes when we try to connect to a remote machine to view the log for the current render for example.
Deadline only stores the host name for a node and not the FQDN (fully qualified domain name).
This means its not possible out of the box to connect to a remote machine as the machine name isn’t recognized in the local network.

We could change the scripts and connect via IP but then we lose the proper machine name for RDP and VNC sessions at the top of the window.

So it feels like this is something deadline should support out of the box.
Feature request:

  • Automatically store the FQDN per node.
  • Automatically fallback to IP address if a connection via FQDN isn’t possible (because the DNS routing is not setup for some reason) in all scripts (to see logs, use RDP, VNC etc.)

Thanks

2

Hey Patrick,

In Deadline 6.2, there will be a Repository Option in the slave settings to have the slaves try and grab the FQDN instead of just the host name.

Cheers,
Ryan

3

Great Ryan. Will it also work for already existing nodes and not just new nodes?

Also what about this one:

  • Automatically fallback to IP address if a connection via FQDN isn’t possible (because the DNS routing is not setup for some reason) in all scripts (to see logs, use RDP, VNC etc.)
    It’s not deadline’s fault if the routing isn’t setup but it would make the deadline user experience so much more reliable if it just works regardless of other issues. The same way that Skype calls just work regardless of firewalls.

Thanks
Patrick

4

There is already an option to use the slave’s IP address instead of its host name for remote control:
thinkboxsoftware.com/deadlin … Settings_2

Could you just use that option?

Cheers,
Ryan

5

It would work but it wouldn’t be nice. The reason is that everybody appreciates seeing a name over an IP address (one of the reasons for the DNS system). Ie when you connect via VNC / RDP its super helpful to see the name of the machine in the window title. Using an IP address should really be a fallback instead of the default.
In our case 90% of the time people connect to a local render node and 10% of the time to a remote one. With a black and white option people have to suffer 90% of the time to enable the usage for 10% of the time.
While using a fallback means it just works smoothly 100% of the time.

It’s subtle but it leads to happy users :slight_smile:

6

Alright, we’ll add this fallback option to the wish list. We won’t have time to add it to 6.2, but we can definitely consider it for a future release.

Cheers,
Ryan

7

Thanks.

In Deadline 6.2, there will be a Repository Option in the slave settings to have the slaves try and grab the FQDN instead of just the host name.
Will it also work for already existing nodes and not just new nodes?

8

Yes. When the slaves launch, they always check their machine name at startup and set it in their state information that you see in the Machine Name column in the slave list. So if this option is enabled, the next time the slave starts up, it will try to find its FQDN instead of just the host name and update it accordingly.

9

Great