AWS Thinkbox Discussion Forums

Troubleshooter's Corner - The machine may not exist on the network

Deadline has always made the assumption that it could connect to Slaves from the Monitor using DNS. Normally, this is a great idea because a machine’s IP address could change, but the host name is fairly consistent.

The problem comes in when you’re working across platforms that do not have compatible multicast DNS mechanisms such as macOS and Windows. You’ll end up seeing errors with remote commands:

could not resolve slave name (MyMachine2) to IP address. The machine may not exist on the network.

An older version of this (or one that can come up with remote command redirection) is this one:

The requested address is not valid in its context 255.255.255.255:35715

The same can happen with remote Slave log viewing:

The slave name provided does not have a machine name or IP address associated with it. The slave may not exist in the repository.

Now, we have a handy feature in Deadline that’s been around awhile that will use the IP address the Slave has registered with the Database. That will be consistent as long as the Slave has run after an IP address change. It’s here in “Configure Repository Options” (similar options exist for Pulse, Balancer, etc):

Usually, enabling this will work. It also takes effect immediately in the current Monitor where the change was made, so testing is quick. Other Monitors that are open when the change is made will have to wait the 10 minutes or be re-opened.

Hopefully that helps a few people (feel free to “me too” in this thread or ask questions).

Now, the problem I have on our side is that we’re not sure if it’s safe to have that on by default. Has anyone had problems that were resolved turning it off?

Privacy | Site terms | Cookie preferences