Version 2.9.1.0

This is an optional release with new features

New features

• Support the Grandine consensus layer client
• Migrate to Loki 3. Caution that if you didn't run ./ethd update sometime between 2024-04-09 and 2024-05-01, this change may require you to stop Loki with ./ethd stop Loki, remove its volume with docker volume rm eth-docker_loki-data (adjust if the directory is not named eth-docker), and start again with ./ethd up.

Bug fixes

• Reth DB conversion check works when there are two Reth on the machine
• ./ethd install doesn't throw an error if Docker is already installed and pkg-config is not

Version 2.9.0.0

This is an optional release with new features

Breaking changes

• Requires Nethermind v1.26.0

New features

• ./ethd prune-nethermind will migrate the Nethermind DB to HalfPath on nodes with 32 GiB of RAM or more
• Add discv5 to Reth

Changes

• Teku default heap to 6g
• Lodestar default max peers to 100
• Use Geth 1.14.0 defaults
• Remove ./ethd prune-geth

Bug fixes

• Do not default Besu to highspec on less than 64 GiB RAM, to avoid OOM errors
• Do not downgrade the .env version
• Fix Loki and prepare for Loki 3
• Check for snap docker earlier

Version 2.8.0.0

This is an optional release with new features

Breaking changes

• Requires Besu 24.3.0

New features

• Supports ./ethd prune-besu for long-running Besu DBs, to one-off prune trie logs

Version 2.7.1.0

This is an optional release with new features and bug fixes

New features

• Supports Reth beta and prompts for resync if coming from Reth alpha
• Update cadvisor to 0.49.1
• Update ethereum-metrics-exporter to 0.23.0

Bug fixes

• Fix HOST_IP in allin1.yml files
• Lodestar dashboard shows more metrics
• Fix ./ethd config not building
• Besu high spec more reliably activates with 64 GiB RAM
• pull_policy: never to avoid extraneous builds

Version 2.7.0.0

This is an optional release with new features and bug fixes

Breaking changes

• Breaks if ext-network.yml was changed. In this case please mv ext-network.yml ~ it out to your home directory, try ./ethd update again, and then nano ext-network.yml and re-apply the changes that were made, likely just the network name as name: rocketpool_net for reverse hybrid.
• Drops support for Docker Compose V1
• Drops support for automatically converting pre-merge (before September 2022) configurations
• Requires Nethermind v1.25.4 or later

New features

• Changed Lighthouse default max peers to 100, to match their 5.0.0 release
• Lodestar flags updated to use the current convention. Thankls @nflaig!

Bug fixes

• Works more consistently on macOS. Thanks to @alindsilva and @toraonion for raising the issue and testing!
• Fixed prometheus-traefik.yml
• Fixed the jwt secret that ./ethd config queries for not being saved. Thanks @victorelec14!

Version 2.6.1.0

This is an optional release with new features and bug fixes

New features

• ./ethd sign-exit all will sign a voluntary exit message for all loaded validators
• ./ethd update --help will print help specific to the update command
• Added a Graffiti length check to ./ethd config
• Changed the CloudFlare DDNS container, as the old one is no longer supported upstream
• Source compile for Geth, Prysm, Erigon and mev-boost uses Go 1.22

Bug fixes

• Updated Discord link. Thanks to @victorelec14 and @ymittal!
• Fixed some typos. Thanks to @cristiantroy!
• ./ethd keys delete fixed
• ./ethd keys prepare-address-change works with Prysm
• Nethermind dashboard fixed
• SSV dashboard fixed
• Fixed the pre-provisioned high memory alert in Grafana

Version 2.6.0.0

This is an optional release with new features and bug fixes

Breaking changes

• Requires Lighthouse 4.6.0 or later

New features

• Supports ./ethd prune-lighthouse to prune Lighthouse state
• Lighthouse VC will broadcast duties to all configured CLs, and uses v3 API for producing blocks
• Besu on machines with 64 GiB of RAM or more will use the high spec flag
• Added Bloxroute relay to default Holesky MEV config
• Prysm uses 70 max peers by default
• ./ethd config offers a VC-only setup on Gnosis Chain. Thanks @haurog!
• Remove Nethermind memory hints
• Add network aliases to CLs, ${NETWORK}-consensus. This can be useful in setups with multiple Eth Docker
  installations on one host, all connected to the same Docker overlay network
• Changed Prometheus yq dependency to version 4
• Update Lodestar beacon max-old-space-size to 8192. Thanks @nflaig!
• Changed Teku default heap size to 5G
• ./ethd keys does some input vaidation on public keys and other user-supplied values

Bug fixes

• Nethermind prune no longer outputs an error message on Nethermind 1.25.0 and later
• Added parameter to support Erigon 2.55 and later

Version 2.5.0.1

This is an optional release with new features and bug fixes

• Breaking for Besu: Requires Besu 23.10.3
• Besu will limit its trie logs, and on fresh sync heal its flat DB for better RPC performance
• Besu source build uses Java 21 runtime (but continues to build with Java 17)
• Additional pre-previsioned Grafana alerts: memory, CPU, out of memory (OOM) kill
• Geth can keep its ancient directory on a separate path, see ANCIENT_DIR in .env
• Logs dashboard works if the directory is not called eth-docker
• Nethermind auto-prune uses 350 GiB threshold on Gnosis Chain again (fixes issue introduced in v2.4.1.0)

Version 2.4.1.0

This is an optional release with new features and bug fixes

• Fix ./ethd keys import with unique passwords. Thanks @shamoya!
• Fix Teku archive sync on Holesky
• Nethermind auto-prune is network aware: Kicks off at 350 GiB mainnet, 50 GiB otherwise
• Web3signer PostgreSQL migration to PG16 during ./ethd update
• Teku source build uses Java 21
• Slightly less naive offline detection for create-withdrawal-change.sh

Version 2.4.0.0

This is an optional release with new features and bug fixes for most users
It is mandatory for users of Teku 23.12.0 who do not wish to use checkpoint sync

• Breaking change: Teku without checkpoint sync (e.g. archive node) uses parameters that require Teku 23.12.0
• Fixed checkpoint sync url query during ./ethd config
• ./ethd keys can set individual Graffiti, as long as the client supports it
• Prometheus instance hard-coded for easier use with dashboards that use instance
• Added a pre-provisioned Grafana alert for disk space
• Renamed Teku and Nimbus -legacy.yml files to -allin1.yml