AWS Thinkbox Discussion Forums

Linux Slave Logs - Firewall ports

Hello All

I recently began testing deadline with nuke on Cent OS.
All works very well, can remotely restart the slave etc after opening firewall ports within firewalld upon reading this:
https://docs.thinkboxsoftware.com/products/deadline/10.0/1_User%20Manual/manual/considerations.html

One thing Im noticing however is trying to look at the logs from that linux node, from deadline monitor on my windows workstation fails. It appears to use random ports??
For eg:
A connection attempt failed because the connected party did not properly respond after a period of time, or established connection failed because connected host has failed to respond 10.20.10.182:37797

So I opened that port, tcp and it now appears to work. But am I wrong to think that its random? It almost seemed to round robin ports for me the other day but I could be incorrect.
Or does pulling logs from slaves always use port 37797??
Thanks

No, you’re absolutely right that it’s random. You can override that with the “remote command port”.

1 Like

Edwin thanks a bunch for the prompt reply!! Now have it set to a static port and all seems well :slight_smile:
Go figure it was staring at me right in the face, just wasn’t looking at the correct spot within deadlines documentation.

All good. I’ve opened a ticket internally here that we just have that on by default and pick a sane number. We do it for the cloud instances so they can work with AWS’ security groups, so why not everyone? :slight_smile:

Ha! Whoa was not expecting this to become an internal ticket but good to hear!
Keep up the great work. Deadline has been a fantastic product to work with, and your guys documentation is very well written, and easy to understand. Take care!

I’ve seen it bite enough people that I’d like it fixed. If you were the first person to bring it up then maybe I’d have let it slide.

Hi we would love to see that link here since our remote log ports change to something random every time we reboot.

But it sends us to 403 forbidden and I can’t find this information anywhere else. Would it be possible for you to show me which file and parameter we can change to make the log port static? We’d really appreciate it

We pulled down the 10.0 documentation, here’s the 10.1 equivalent. If you don’t want to hard-code the value for every worker, you can set a farm-wide override in the repository configuration. Just be aware that if you’re running multiple Worker instances on a single machine that will result in a port collision.

Hello @Chris_Spencer

We have changed the terminology from Slave to Worker. Here is the updated link: Worker Configuration — Deadline 10.2.0.10 documentation

Hey @zainali What field exactly should one set?

Changing Override Remote Command Port does not really solve the issue as Spot Event plugin creates new EC2 instances in my case so overriding the port each time does not work long term.

What I mean by does not really solve the issue is that Monitor keeps trying to connect to random port for me. Should I restart Spot Event plugin and shutdown all instances first?

@Petr I see what you mean. I think you need to set this setting globally from the Monitor> Tools> Configure Repo Options> Worker Settings>


Check this ^ option and set a port override. If you set this globally more than one Worker on a single machine may run in to connection issues because of the port conflict.

Hi @zainali I did that and it helped. However, the problem with EC2 instances is incorrect public IP address.

For instance, Monitor picks the internal lPv4 address instead of using public IPv4
ec2-3-235-194-XX.compute-1.amazonaws.com to connect to the Worker log.

Can I resolve it somehow without manually adding Worker machine to C:\System32\Windows\drivers\etc\hosts file?
image

What solution are you using? Monitor does not connect with the Public IP usually. I think this is a firewall issue if Monitor is not able to stream log. I see the port used was a random port. Did you use static port instead in the Worker properties override (as I mentioned in my original response)?

If there is a static port and Worker is not still able to connect, then check the network traffic on the port used for it. You will need to use "telnet -tulnp }| grep " to check what is going on on that port. Please

Privacy | Site terms | Cookie preferences