multi-region devnet deploys demo

[meaghanfitzgerald]

Use Cases:

• private testing and load testing in early phase with the partner so they don’t need to test in public test net
• Large scale load testing of avalanchego or custom VM without public test net
• Run production ready validators using aws
• Distributed load generator for evm performance testing
• Multi region deployment to simulate real world network topology

ava-labs/avalanche-ops#310

You can either specify an explicit list of regions to use or simply provide the number of regions you'd like it to auto-select during deployment. avalanche-ops takes care of assigning nodes to regions for you.

We used this tool to automatically spin up 100 nodes in ~18 minutes (and tear it all down In 2 minutes). We'll use this for the upcoming HyperSDK devnets but have been using it for AvalancheGo optimization work in the meantime.

Instructions:
First, you need access to our experimental dev account (if you haven't, you can ask our IT team to give you access)

Second, you can download the avalancheup-aws binary from here https://github.com/ava-labs/avalanche-ops/releases/tag/latest

Third, you can run the commands as shown in the hyper sdk readme or for example,

avalancheup-aws default-spec \
--arch-type amd64 \
--os-type ubuntu20.04 \
--anchor-nodes 4 \
--non-anchor-nodes 4 \
--auto-regions 3 \
--instance-mode=on-demand \
--instance-size=2xlarge \
--ip-mode=elastic \
--network-name custom \
--keys-to-generate 10

[yashdev9274]

Hey @meaghanfitzgerald is this issue open to all, I would love to know about this issue to work on it

[meaghanfitzgerald]

gm @yashdev9274 yes, anyone is welcome to work on this demo. All the information I have been given is included in the issue description, but if you have any questions you can reach out to me or @gyuho

[DevRelMalima]

GM @meaghanfitzgerald is this issue still open for contribution, if yes I will like to participate