New to SPs - Backup Block Producer, Server sizing

Hello,

We are planning to run and operate stake pools soon, and before doing so, we are in the stage of learning about it through the various posts and videos (the community is amazing in sharing and helping each other, amazed!). So as we read and learn, questions popup, so here are some questions:

Backup/Standby Block Producers

  1. To achieve extremely high availability and reliable pools, we want to take this into consideration from the beginning. many people mentioned to have a backup/standby BP in case of maintenance/upgrade of the primary one. To setup a backup bp, we build it as if we build the primany one? any specific instruction for this?

  2. Can you run 2 BPs in active/active setup and load balance them?

Server Infrastructure

  1. We are planning to use AWS (we know, it’s not as easy as the other hosts, but we want to learn the AWS as well in this journey). Can we have the relays in different availability zones? Will this (the latency) be an issue between the relays and the BP?
  2. In terms of storage, any recommendation in terms of IOPS?
  3. Is it better to use a different partition/drive for the cardano-node or its just fine to have it on the OS drive?

AK

Hello,

Welcome aboard!
Nice to see new people joinging!
If u will need help this is the right llace to ask!

Now

  1. To setup a backup bp, we build it as if we build the primany one? any specific instruction for this?

Nope, all nodes have the same configuration - it will differ only few scripts and the way how are started (as a Producer or as a Relay)

  1. Can you run 2 BPs in active/active setup and load balance them?

Nope, u can’t have 2 BPs running as a Producer same time; u have 2 possibility :

  • start the bkp BP manually when the main fail
  • create scripts which will start automatically the node as a Producer when the main BP will fail
  1. We are planning to use AWS (we know, it’s not as easy as the other hosts, but we want to learn the AWS as well in this journey). Can we have the relays in different availability zones? Will this (the latency) be an issue between the relays and the BP?

Yes, will be fine; u can run BP and 1 relay same place and another relay/relays different area

  1. In terms of storage, any recommendation in terms of IOPS?

Minimum 100G (greater =better)

  1. Is it better to use a different partition/drive for the cardano-node or its just fine to have it on the OS drive?

It will be ok both cases, it’s up to u how u will configure the servers

Cheers,

2 Likes

Thanks Alex.

Before starting with mainnet, can you refer us to a guide for the testnet? This way we will experience it hands on prior to mainnet setup.

Cheers,
AK

100G for BP or Relay or both?

All nodes will have the same DB so… for all

@AK12345 @Alexd1985 - Jumping in on this convo lol. I’m trying to setup a standby BP node as well to minimize downtime when the main node is undergoing maintenance. My friend and I are fresh on the scene, our pool has been up for just two days (ticker “MPOT”). When it comes to all of the keys and cert files, is it sufficient enough to have it synced to the blockchain, and then just copy over the kes.skey, vrf.skey, and node.cert files from the main BP node in order to run the cardano-node run command on the standby node?

These three files I mentioned seem to be the only difference in parameters from running as a relay and running as a BP node - Generating Keys for the Block-producing Node - CoinCashew

This is true, only these 3 files… but also u can upload them anytime (don’t forget to re-upload each time when u will rotate the KES) and also configure the startup script for BP mode and when:

  • u will want to start as a BP run startBlockProducingNode.sh
  • as a Relay run startRelayNode.sh

Also u can find scripts which can automatically switch between Producers on github

Cheers,

1 Like