Hay
I am working on a small render farm deployment script. This works but when I create two nodes they show up with matching UIDs in the logs.
Also when using the cloud option, where if I logged in to the head node, it only comes up with the last one that was created?
advise is welcome on this
Hello Nigel, :) ok the issue here is that AWS VMs will all return the same hardware UUID because they are virtual machines with exactly the same image running on them.
We specify a command line argument for the headless server that might help you to circumvent this.
This argument is given the option name "-ak" and the expected value is a UUID formatted string (see https://docs.python.org/3/library/uuid.html which explains the format, we use the python uuid library to generate these uuids). This overrides the UUID that the headless server would normally get via introspection of the hardware it is running on, which in your case, with AWS, is virtual hardware, and as a result, no matter how many nodes you start, will all return the same uuid.
This argument will allow you to generate your own uuid and pass it to the headless server when starting it with the command line. The value you pass will then be used as that node's uuid.
As for the cloud issue you have, are you able to post a screenshot? Not exactly sure what you mean in your description of the problem, I'd love to help, but at this stage, I am a bit confused! Sorry about that :)
James
Just an update
both issues have been fixed :) i used https://www.thegeekdiary.com/centos-rhel-7-how-to-change-the-machine-id/
to change the hardware id on redhat/centos boxs
just need to look at how to remove old nodes from my list
Thanks i had an felling it was an ID but was not sure how it was generated,
I give this a go i think the two issues are linked, so i try this first