-
A Wake-on-LAN packet can be almost anything, it just has to have the target computer MAC address in it, 16 times, and somehow get to the off computer.
As the computer is off you can't assume a local router will have an ARP entry for it and you can't assume the switches will have seen any data from it recently so will not have a forwarding entry for it's MAC. Hence usually sending the magic packets to a broadcast or multicast address that will be flooded everywhere including the target.
If you have a computer on the same broadcast domain as the render nodes then you can send the magic packets with a utility directly to the LAN broadcast address. You will need to send (at least) one packet for each computer you want to wake.
The Wikipedia article lists a bunch of example software that will send the packets or a bit of Python code if that is your thing. https://en.wikipedia.org/wiki/Wake-on-LAN#Creating_and_sending_the_magic_packet
You are reading a single comment by @jellybaby and its replies.
Click here to read the full conversation.
jellybaby
Polygon
I've lost you a little bit, I thought my issue was less sending the packets, but rather them finding their way through the network to the intended recipients. If we assume that 1 computer is always on (the computer I'm remoting in to) then the issue is getting our render nodes to boot when no one is in the office. Assuming network card and bios are set up correctly on those machines what would that solution look like? I thought the magic packets had to be sent to broadcast on port 9. I can't set up multiple IPs as the recipient of port 9 it seems.
In case you can't tell I'm fairly networking illiterate.