Grant: Node Pilot (1-click deployment and validation)

The conversation here interestingly shifted towards validator role quite a bit. In looking closer at the proposal, I believe the proposal is NOT for running validators.

It’s to operate a non-validating full node using a 1-click approach, and yet be able to earn POKT rewards for serving RPC traffic.

I believe that’s the primary purpose behind this. Making it a validator node becomes secondary. Having validator node running experiences will help though.

4 Likes

Ah, ok.

IMHO Validator creation could work with docker.
It could be docker-compose.yml project description opensourced, located and managed in Github.

But all sensitive info ( BLS keys, BLS Keys passwords, Validator’s Wallet) should remain in control of node operator. this files could be just Mounted to the docker container to appropriate folders, from local Node folders… That are located there by owner itself after git clone of Docker project.

1 Like

This proposal is bascially running on a Docker node. I half built it earlier in the year but then with EasyNode it seemed less likely… Plus Docker adds extra complexity that I don’t believe is required and of course extra maintenance, support etc.

1 Like

Ah, ok, sure. If we will skip Validator Node creation part here, team would benefit from such automation for sure.

1 Like

Maybe @shane can update the proposal to reflect that as that indeed could prove useful but at present the objectives encompasses all node setups.

  • Harmony Validators can “double dip” in the native ONE rewards and Pocket rewards.

Maybe there is some misalignment here.

Or maybe split the proposal into 2 if there is scope for both.

Auto POKT / Harmony Full Node RPC would be good as I know there has been some technical barriers for setting up properly.

For those who this may apply to, Happy Thanksgiving!

Don’t have much time today but I wanted to quickly jump this growing discussion.

[Disclaimer - I LOVE GUIs and LOVE Docker] :wink:

Really impressive products. I’ll dive into them deeper a bit later, but the documentation is very impressive.

Yes. NP is intentionally designed to allow nodes to be portable. You can easily move a validator that was started in NP to your own Docker instance. There is zero platform lock-in and our deployment engine ensures everything is transparent. Our file structure and configs are easily understandable and designed for portability.

We do not manage the relationship between delegators and validators. Delegators should choose wisely who they want to delegate with… we have no involvement. We just make tools that streamlines how node runners can deploy and interface with their validator, but the responsibility of being a good validator to their delegators is still their responsibility.

What I have seen in most networks is eventually the majority of validators are run by custodial services. There are always a portion that are independent node runners, but most networks have the majority of validators centralized into only a few providers. The worry is then “well how do we break up the providers?”. By then you are most likely too late.

Node Launcher is ultimately the middle ground, because it enables anyone to make validator deployment applications seamlessly through JS. Node Pilot acts as the first use-case of it, but we ultimately want others to build their own and democratize node running with many options.

Yeah, that is the risk with custodial services. NP is self-hosting, the polar opposite.

We plan to do with Harmony what we’ve already proven with POKT. Once you deploy a POKT node and you have met the threshold to be a validator (have enough POKT, connected chains, etc) then we have a “Stake Validate” which does the validator process automatically once clicked. We are planning to have the same UX for Harmony validators.

I may have under-represented Node Launcher and its importance in all of this. Node Launcher is the secret sauce and the real tech we are pushing… as it enables programmatic deployment/management of Harmony containers through JS. We originally launched NP back in April where we did just hard-code docker files for nodes. We quickly realized once we started stacking different types of nodes that we wanted to standardize how nodes can be deployed and updated. So instead of managing hard-coded docker files, we have a programmatic platform that we just add integrations into.

I’d encourage the more tech-inclined to check out the Github for Node Launcher and perhaps our 7min introduction video: Node Launcher - A TypeScript Library Introduction (POKT Community Call) - YouTube

Feel free to check out the documentation for deploying a POKT node the traditional way. For the IT inclined, it’s absolutely doable without NP, but the Pocket community kicked off this project because custodial service providers were becoming the norm and majority of the network.

I may have missed some important points, but I tried to get in as many comments from as many individuals as I could before focusing on the holiday. Feel free to keep the convo going and high-light areas I can provide better clarity :+1:

:turkey:

2 Likes

One more point I missed: This expands what kind of applications can be built for nodes and validators. Others could take Node Launcher and build a Harmony branded application that functions similar to Node Pilot. Or an existing application, say a wallet, could add validator abilities with a few lines of JS code. The more tools for the community/ecosystem the better IMO.

Harmony will probably also have other community-driven services that will likely rise up and require some kind of node. Like with other platforms, common examples are decentralized bridges or privacy relay networks. On other networks, those are complex deployments going the traditional route, but they can be standardized and launchable with Node Launcher, making any app with Node Launcher able to easy add new services. If it’s added to NL, then it will also be compatible with NP and any others.

Through our experience with trying to develop NP, we think a standardized, programmatic deployment platform will scale and generate more adoption than standard docker file setups. We may be the crazy ones though :joy:

1 Like

Thank you for taking the time for a detailed reply… I have checked out the github and like the transparency.

I will hopefully have a deeper look over the next few days and I wish you happy holidays :slight_smile:

1 Like

Thank you for taking the time.

1 Like

Great set of discussions here. Fits the bounty category in Apply for Grants or DAO

:tada: Approved for the $30K to build initial integrations and docs, and $21.5K to operate validators in multiple shards for 18 months with testing, maintenance and support.

Really looking forward to seeing this go live and be in the hands of new and existing node runners!

1 Like

We very excited to make this happen!! Q1 for node runners is going to be :fire:

We will keep folks updated as things progress. In a few days we are releasing one of our most feature rich updates yet for Node Launcher and Node Pilot, most notably enabling a new way to seamlessly update. Things are moving fast and are excited to dig into Harmony :slightly_smiling_face:

5 Likes

Hi,

In order to know account balance in older blocks I need to have access to an archieval or explorer node.
(Validators nodes has not this info)

So, I am not interesting in deploy one by myself, but I am wondering if you offer a way to have access to one explorer node to send a few queries.

Thanks :slight_smile:

POKT plans to add archival access, so that will be an option in the future.

Node Launcher could be used by other to setup new node services, but we do not provide RPC services.

2 Likes