Pre-Dev-Update Thread! Yay! :D

Development Updates
5353

Tut, tut, I see you aren’t following the github closely enough and shall notify the re-education department.

3 Likes
5354

Guilty as charged. I will try to do better.

To be fair though it is a full time job to stay on top of it!

5 Likes
5355 
info!("New peer added to routing table: {peer:?}, now we have #{} connected peers", self.connected_peers);

:pray: @roland

11 Likes
5356

We’ve also added a log marker for it, PeersInRoutingTable(usize). You can maybe grep for this instead, as this should be less prone to accidental deletion.

6 Likes
Vdash - dashboard for Autonomi nodes
5357

Nice!

10 Likes
5358

I recon it must be about time for a testnet :slight_smile:

9 Likes
5359

An interesting issue, defines a new binary ‘node-manager’ and commands which will be used to manage nodes as services, as an alternative to running then manually.

8 Likes
5360

Quotes are back??

4 Likes
5361

I was just about to post that! Maybe they reduce the messaging overhead as nodes commit to a price and need to store? No constant re-checking with every failed chunk?

4 Likes
5362

They do that, but they’re not back in the prev sense. See the updoot for more deets

7 Likes
5363

Has anyone tried this?

Looks like a awesome bit of kit @chriso :clap:

12 Likes
5364

Hey Josh, thanks. It’s still very much under active development and subject to change, literally on a daily basis. The short-term aim is to use it for upgrading nodes, which would enable us to push out changes but keep our testnets running.

Of course if you decide to use it, any feedback would be welcome. I might recommend just running it on VMs for the moment. It installs services on the machines where it runs, and how it does so can be subject to change, so it could potentially leave your machine in a state where you would need to manually clean things up.

14 Likes
5365

That is a game changer.

Thanks for the tip on using a VM I’ll definitely be kicking the wheels.

16 Likes
5366

Ledger hardware support being worked on:

14 Likes
5367

You can tell things are hotting up when the forum is buzzing and people are too busy to post here for twelve days!

Meanwhile, here’s a boring testing PR being used to sneak out the following task at the bottom of a to-do list :mag: :male_detective::

client/node API documentation and testing

12 Likes
5368

Will be next update huge? :smiley:

1 Like
5369

Testnets seem remarkably stable these days so we might see more stress testing, such as with churn, but there’s already a new issue that I think needs attention.

At the moment pricing seems to cause great disparities in node earnings, which will be bad if it causes lots of churn while people restart low earning nodes to boost earnings.

So I think we’re going to need quite a few tests to figure out how to price storage in a way that manages supply without causing unwanted inefficiencies such as churn caused by node ID fishing.

10 Likes
5370

I wonder if the issue with big uploads is likely to be addressed. So far I don’t think anyone has been able to upload that 36gb LLM if I remember correctly?

What are the main issues that the team are aiming to sort in the next tests?

Are they;

  1. Pricing model / tokenomics to promote more even earning between nodes
  2. Big upload issues

Are there any other big issues?

Certainly seems like the tests are doing great and seems like not far to go, though the last few issues can be the toughest! Looking forward to the next tests whatever they may bring :slight_smile:

9 Likes
5371

Correct. I have it on a cloud snapshot, ready to try again on the next testnet. My max upload from home is ~4Mb/min so I need to run this from the cloud to see what the real bottleneck is.

6 Likes
5372

I will try again on the next test as well. I have fairly fast upload speeds on my connection, but I expect that isn’t the bottleneck with uploads to current testnets.

5 Likes