Introducing [RAOOL] - Raspberry stakepool


Successfully minted blocks :slight_smile:.
As you see the Pi isn’t overwhelmed at all with the task. Just building and syncing the node took the Pi a little bit of effort.
Does someone know how to add the Pool vKey at ff.pooltool.io ?
Pool Id: cf5091f38cb4821e582c6833aff9026383a5ac5ecb38bcc660e71860ead6a121

3 Likes

hi @alessandro
can you share the details of how you cross-builded cabal for the ras pi?
I just got my ras pi 4 and eager to try it out

1 Like

Hey, so basically you need Ubuntu 20.04.
You need to install the haskell-platform:

sudo apt install haskell-platform

now you have ghc8.6.5, but cabal2.4, then:

wget http://hackage.haskell.org/package/cabal-install-3.0.0.0/cabal-install-3.0.0.0.tar.gz
tar -xf cabal-install-3.0.0.0.tar.gz
cd cabal-install-3.0.0.0
cabal update
cabal install

This will generate a cabal 3.0 binary for aarch64. Now you need to remove the cabal2.4 from the PATH and add the cabal3.0 binary to .local/bin. The rest will be the same like building a normal node from source.
I will probably create some scripts and tutorials on my github.

4 Likes

@alessandro thanks a million alessandro. I’ll give it a try and let you guys know how it went. looking forward to the tutorial in github :slight_smile:

No problem :slight_smile:

1 Like

Hi @alessandro Do you confirm that HTN run smoothly on a RasPi 4 8Gb ?

Yes I don’t have any problems so far and it’s always in sync. I think the 4GB version could also work without problems.

1 Like

Thanks for your reply, that’s a really good news!

I’ll give it a try on mine.

Hi to all raspbians,

I have a pi4 4gb as well and i want to use it just for node. My question is…can i run 2 nodes on same Pi ?

1 Like

Hey I just created a new Post Guide Raspberry Pi 4 HTN Setup [Pi Pool]. I just created a guide to how to setup the HTN on Raspberry Pi. I hope it can help some of you.

4 Likes

thank you alessandro

1 Like

hello Alessandro,

hope all is well!

i am looking to set up two raspberry Pi’s, do you have a guide or know one i can follow to get started, please?

kind regards
R

Hey, yes you can check this out: GitHub - alessandrokonrad/Pi-Pool: Cardano Stakepool on Raspberry Pi

1 Like

You can also install Docker on the Pi and then simply do …

docker run --detach \
    --name=relay \
    -p 3001:3001 \
    -v shelley-data:/opt/cardano/data \
    nessusio/cardano run

docker logs -f relay

The image also supports built-in …

  • gLiveView
  • topologyUpdater
  • slot leader logs

The details of which are here

Correct spelling for the era is Shelley, but I don’t run a pool, maybe the coders got it wrong, making this line correct…

This is entirely my mistake - thanks for pointing it out. I’ll fix the docs.

1 Like

Hey there,

This post was very helpful, thank you!
I’m creating a test pool in aws (since i’m most comfortable there) but I would love to build a pi setup, maybe I can pick your brain if I get stuck on something. :slight_smile:

Thanks,
Matt

Sure you can. If you use AWS you might as well use an arm64 image from them. They have CentOS8 on arm64 for example. In that way you can build docker images on AWS that will then run on your Pi as well.

Unfortunately, the upstream Nix build does currently not work on arm64 because of [BUG] - cardano-node build fails with “Segmentation fault” on aarch64

For the nessusio/cardano images I have a convoluted workaround that builds the arm64 derivation through Docker (instead of Nix). Once I have the the binaries for the node & cli a normal Nix build can continue.

If you like this work, you may want to endorse this [CIP] - Provide high quality multiarch docker image and k8s support. The PRs that I have so far are here

Hello there!

Here is an update of the pool situation, as this is its introduction page after all :smiley:

The first post mentions that I used one RPi 4GB. This was true only during ITN, back in April 2020.
Since Shelley, I’ve been using two RPi 8GB.

Building this pool with my two Raspberry Pi has been a wonderful journey.
But as improvements on the Cardano protocol keep coming, I find it hard to rely on them solely. I’ve noticed that transitions between epochs often result in CPU/RAM overload, giving a hard time for the Pi, especially the one that hosts two relays and the block-producer.
This situation made me miss a block that was scheduled very early in epoch 251, I have to be honest.

Thanks to IOG delegation program that started in february 2021, I’ve been able to produce 11 blocks.
So, I decided to use part of the received rewards to buy a barebone server with the following specs :

CPU: Intel i5-8259U Quad-Core 2.3GHz
RAM: Corsair 16 GB DDR4 2400 MHz
SSD: Corsair MP510 960GB

This newcomer is dedicated to be the block-producer, nothing more. It runs since mid-epoch 251, the transition with epoch 252 was totally smooth and it has produced its first block this morning.
No relay runs on it, even if it’s powerful enough. I keep this tasks for the RPi. The one that used to run two relays and the block-producer feels better now ^^
Another bonus that comes with this new architecture is the extra layer of security as the block-producer now sits behind the relays, not next to one as it used to be.

2 Likes