With Deadline 10.0.17.5, the remote control loop back to the machine sending the request (using ip, but my config use machineName).
So all remote commands I send from one machine to other machine loop back and are excecuted by the machine sending the command (restart machine/slave, stop/restart Machine/slave after task etc…).
I personally use Machine Name instead of IP, but when sending requests via monitor I notice that the Machine name is, with D10.0.17.5, inside remote commands pannel, the ip of the sending machine.
Would you be willing to try 10.0.16? We made some changes to remote commands (though it should only be for WoL) and I want to try to limit the search space down to only one release instead of 9.
Historically, commands being looped back on itself was due to DNS issues where the machine would try looking up a remote machine and it would end up returning the wrong IP. Are you using hostname / IP address overrides?
Hi, this behaviour only apply when using 10.0.17.5, all previous version, work just fine, but not this one.
All my slaves have a static IP distributed by the router, always the same fixed ip
I use hostname override.
=> If I send to computer B from computer A a cmd (start slave, restart machine etc):
-> with previous version, the machine name was Computer B name
-> with 10.0.17.5 , the machine name is ip of Computer A and Computer A will excecute the command.
So, here what I tried, each time on 2 different machines sending to the other one and 3 deadline version (10.0.16.6 | 10.0.17.4 | 10.0.17.5)
For all test :
Repo settings :
-Use FQDN : ON
Slave is up and running.
Test #1 and #2 :
Repo settings : Use slave’sP for remote control ON (#1) or OFF (#2)
slave properties : Host Name override was used
Test #3 and #4 :
Repo settings : Use slave’s IP for remote control ON (#3) or OFF (#4)
slave properties : IP address override was used
Test #5 and #6 :
Repo settings : Use slave IP’s for remote control ON (#5) or OFF (#6)
slave properties : HOSTName/IP address override not set (blank)
I did this tests, with local network, set as double stack (IPV4 & IPV6), and IPV4 only for each version sending command to the other machine on both machine tested
both previous version (10.0.16.x and 10.0.17.4), no matter the computer used, work perfectly fine, as expected (when Hostname is set, remote command return me the hostname, but when this is ip override, it return me the ip of the target machine, and the command is excecuted on the target machine)
With 10.0.17.5, that another sotry, in this scenario no matter the test, 'remote commands, return me ONLY the ip (V4 or V6 depending of the local network stacking), of the sending machine, and the command is executed on the sending machine…
Thanks for the thorough testing! We’ll take that in and look over the code. We’ll fold it into the investigating we’re doing for this thread: forums.thinkboxsoftware.com/vie … 11&t=16186
