Update 29 August, 2024

Rounding up the news this week, on Tuesday we published a new version of the node so if you’re a beta tester make sure you upgrade ASAP.

Headline improvements in the latest version (2024.08.2.3) are as follows:

  • Updated Launchpad with several new features and controls
  • Connectivity improvements in node-manager, where the start command now waits for one node to connect before starting another.
  • Major improvements to the way home nodes connect. There were a few issues here, but long story short, you should see a big difference to the flow of nanos now.

The full changelog is here.

We also fixed a problem with the uploaders, so the network is being populated with data once again.

The over-quoting issue has been fixed, so cancel the Lambo, there’ll be no more overnight nano millionaires.

In the works are new metrics and settings for the Launchpad to enable users to check the health of and restart individual nodes, rather than the current all-or-nothing.

You’ll also be able to select custom ports and control communications protocols from within the Launchpad, until the time when libp2p solve their AutoNAT issues and we don’t have to any more.

Oh, and a first version of the API will be published very shortly. We’re working towards a simplified replacement for the old sn_client API. Consolidating code which has bled between that and the cli into something which will make it easier for developers to start to get to work on their apps, on which ultimately the whole Network depends. Obviously it will be basic, but at least you will be able to see what we’re up to and offer feedback.

Beta news

For those who missed Discord Stages on Tuesday, we’ve made some small adjustments to the Beta programme that reflect our current plans as we approach launch:

Wave 3 will start once we reach 500 participants rather than 1000 (and accordingly, half of the original prize pool). Any earnings, including those from boosts, will contribute to the cumulative balance for the opening week. Wave 3 will be the last wave, as the beta has fulfilled its purpose, and focus will shift to other testing and incentive programmes. Wave 3 will have a fixed end date, no later than Friday, October 10th.

Prize Pool Adjustments: Starting Wave 3 will increase the total prize pool by 50% for Wave 2 and 25% for Wave 1, with Wave 1 participants already having their pool rewards doubled.

Post-Wave 1 Incentives: When Wave 1 ends in two weeks (around week 10), participants can opt into an incentivised token testing programme starting the week of September 29th.

After the end of their respective waves, participants from Wave 2 and 3 will also have opportunities to join the incentivised token testing programme. Further incentives will be available at the network’s official launch.

Team news

To help us in the push to the finishing line, we’re pleased to welcome a new Rust dev Ermine Jose to the team. :tada:

General progress

Back from holiday, a refreshed @joshuef is straight back into the thick of things.

@qi_ma worked on a fix for a register bug to ensure they are now merged on the network side where possible. He helped with documenting the sn_registers API readme, and also investigated the stalling uploaders - which meant that no new data was being added to the network - and implemented a fix there.

@bzee has been putting finishing touches to the file API as well as attempting to use cursor.sh, the AI coding tool, on NixOS. He likes a challenge. He’s also the lead on the main API and is implementing registers for updating directories, etc.

@dirvine has been thinking about our data types and the API, and how we can get stuff out to third-party devs as quickly as possible.

@anselme pushes on with the spend refactor, which should be merged very soon.

Over on Discord @rusty.spork continues to answer questions and collect information that deEEP are asking about, as well as generally manning the Discord fort.

Meanwhile @Chriso has been the main man on the new release, together with @shu, who was verifying that the latest release was good to go, using the new automated testing and release procedures we now have in place.

@roland has been automating the testing of home nodes as well as taking Ermine Jose through the codebase.

New dude Ermine Jose has dived straight in, submitting a new PR for WAN nightly test, familiarising himself with the code base and executing a basic build.

@mazzi has been testing the Connection Mode Custom Ports coming to a Launchpad near you very soon. He’s working on status screens when changing states and an error popup.

Meanwhile, @Mick.vandijke continues to work on changes in the wallet capabilities. Again more on that soon.

And @jimcollinson is investigating the possibility AI to enhance developer documentation, and our project knowledge sharing capabilities.

If anyone has any experience in documentation tools that are usefully incorporating RAG that can ingest code and other documents and prioritise the newest materials, let us know!

55 Likes

First !!! Super happy to read the API is coming soon. Let the real fun begin :smirk:

27 Likes

Second - beauty! Nice update team. Welcome Ermine Jose! You’ve jumped into an :ant: nest of busy activity - have fun and enjoy yourself :wink:

As usual, thanks to the team and all the keen supporters :heart:

Cheers :beers:

23 Likes

Third :slight_smile: yaaay !

Only the truly venal amongst you would have noticed this was a very very quick post to snatch the final podium place :slight_smile:

Good work from all concerned, as ever we thank you.
Still many questions around just how this is going to be ready for an October release, but we live in hope and have faith in the team.

22 Likes

Wave 1 should also get 50% larger pool as wave 2. Wave 1 still suffers a large reduction in rewards, as a relative small % managed to earn nanos in wave 2 and those who don’t only get rewarded 50 per week, then there is not that much difference between wave 1 and 2 for rewards. Current switch that wave 1 only gets 25% larger pool is not fair in relation to the rules that was set from the beginning.

8 Likes

Your reward is being allowed to test this fantastic project.
You should feel honoured and extremely humble that this is so.

Instead, you whine.

16 Likes

:partying_face: :boom: :man_dancing:

18 Likes

I’m curious what are those issues?

7 Likes

Also could the BambooGarden fund be used to accelerate resolving these issues?

5 Likes

yeah…didn’t see that in my excitement.

Great job @maidsafe!!!

14 Likes

Thanks so much to the entire Autonomi team for all of your hard work! :man_factory_worker: :muscle: :man_factory_worker:

And good luck to everyone trying to get more nanos! :four_leaf_clover:

10 Likes

Dunno, maybe it is already being tapped to pay for the new guys @mick.vandijke @mazzi and Ermine Jose?
Welcome guys, looking forward to your contributions :slight_smile:

9 Likes

fantastic job to all the team keep up the great work :slight_smile:

here are some more uploads so anyone who wants to test a download can :slight_smile:

"An_Introduction_To_The_SafeNetwork.mp4" a6cced483052f97dbe20b5e12f9fa1d6a55aa44e35c99067c89e47401fb34de5
"AnarchyInTheSouthside.mp3" c0893739a688b739aac5b75ac845b6c428c21b73dc782352bb1ad529495a87af
"BegBlag.mp3" 3509bad03dc869dec883c7b44662c3503d2517fa9e828bb64f4dbe719d3837bf
"Best_CD_Ever.mp3" 560678b8585c48bc67b547b3dfc517eaac1a355323670ab6ea0b4e79877238ed
"autonomi.mp4" 20a4d80fc8e6c60250e940f9797dc28524e0d093ab07e4016027ff517c44c926

also latest release of the nodes is fantastic I am seeing much better performance all round and nothing is catching fire :slight_smile:

iv never seen so much green on my dashboard :slight_smile:

@JimCollinson i ran 600 nodes unupgraded for 24 hours no loto chunks were seen they kept earning normal amounts now iv upgraded them also.

13 Likes

Just one observation @joshuef @dirvine… I’m controlling my node count resource dependant… So a larger and stronger machine automatically runs more nodes than the little 2 core machine…

After this upgrade some machines where I set a very high node limit (powerful dedicated machines where I simply didn’t know a reasonable limit) suddenly ran twice the nodes as they did before the upgrade… And at the same time have seem to be running into connection limitations so it really became annoying to see how healthy they are… Sometimes seconds delay when typing… I did then reduce the node count significantly… But when looking at the earnings they did indeed earn proportionally to the node count… So I’m not sure if they really weren’t a burdon to the network and it was just my ssh session that became annoying… But hard to believe I was still in the area of a “healthy node” there…
… I would think we are still keeping too many nodes in the network that are slowing down everything…

8 Likes

yep Im seeing the same. Looks like I will run out of RAM first on 128GB 24-core box. 440 nodes purring away at load_avg 38-40 - but I have 800MB free RAM with 26 of 64GB swap used.

same here. I was trying to use lnav to grep across the logs but I had to give up. I couldnt enter a filter string as it took so long to echo back my input. OK once I closed lnav though. Likely lnav itself was just too much extra load.

3 Likes

Considering that this week ending Today, Friday, will be week 11 I think you might have this wrong. Wave 1 will be finishing on week 12 as it always was going to. Assuming 2 week approx from the events. (1 1/2 weeks really)

Welcome Ermine Jose to team from the community and myself. Hope your stay with the team will be enjoyable and fruitful

Make sure its not your nodes using it. Nothing worse than a protocol getting swapped out or its cache. Thats when a disk cache is no longer saving time or disk I/O

2 Likes

Im killing nodes to get the RAM down
then sudo swapoff -a
edit /etc/fstab to 8GB swap,
echo 10 > /proc/sys/vm/swappiness
and turn the swap back on

Does that sound right? Swap is generally set and forget for me…

3 Likes

With that much RAM for running nodes I’d leave swap off totally. You have no need and let nodes die from oom if they try to use swap

5 Likes

Im testing one of @aatonnomicc script that stops/starts nodes based on load/RAM used etc
Lets see how that goes with no swapfile.

I have these SSD disks

sudo fdisk -l
Disk /dev/nvme0n1: 1.75 TiB, 1920383410176 bytes, 3750748848 sectors
Disk model: SAMSUNG MZQL21T9HCJR-00A07              
Units: sectors of 1 * 512 = 512 bytes
Sector size (logical/physical): 512 bytes / 4096 bytes
I/O size (minimum/optimal): 131072 bytes / 131072 bytes
Disklabel type: dos
Disk identifier: 0x06e42416

Device         Boot    Start        End    Sectors  Size Id Type
/dev/nvme0n1p1          2048    8390655    8388608    4G fd Linux raid autodetect
/dev/nvme0n1p2       8390656   10487807    2097152    1G fd Linux raid autodetect
/dev/nvme0n1p3      10487808 3750746799 3740258992  1.7T fd Linux raid autodetect


Disk /dev/nvme1n1: 1.75 TiB, 1920383410176 bytes, 3750748848 sectors
Disk model: SAMSUNG MZQL21T9HCJR-00A07              
Units: sectors of 1 * 512 = 512 bytes
Sector size (logical/physical): 512 bytes / 4096 bytes
I/O size (minimum/optimal): 131072 bytes / 131072 bytes
Disklabel type: dos
Disk identifier: 0x8aa72153

Device         Boot    Start        End    Sectors  Size Id Type
/dev/nvme1n1p1          2048    8390655    8388608    4G fd Linux raid autodetect
/dev/nvme1n1p2       8390656   10487807    2097152    1G fd Linux raid autodetect
/dev/nvme1n1p3      10487808 3750746799 3740258992  1.7T fd Linux raid autodetect

I’ll bow to your superior wisdom but my thinking was if I have to swap, these disks are where Id be least uncomfortable about swapping to.

3 Likes