Releases: rocket-pool/smartnode-install
v1.12.0
This is v1.12.0 of the Smart Node.
It is a recommended update for all users as it contains important new features and bug fixes.
Client Updates
- Lighthouse updated to v5.1.2 - includes a block lookup efficiency fix, a fix to stop propagation of slashings for attesters who have already been slashed, a fix to the
--disable-upnp
flag, and a couple of metric fixes; - Teku updated to v24.3.1 - includes improvements to CPU and NETWORK issues. Also introduces Validator Slashing Prevention feature.
- Prysm updated to v5.0.2 - This release has many optimizations, UX improvements, and bug fixes.
- Nimbus updated to v24.3.0 - includes additional beacon API support and resilience to suboptimal network conditions.
- Reth updated to v0.2.0-beta.4 - Improved performance and bug fixes. (Testnet only)
Note that some clients, like Teku, Prysm and Nimbus, started to prioritize locally produced blocks unless MEV-Boost provides a 10% higher value. We are supporting this decision from the client teams and the Smart Node will use these default values.
Smart Node Changes
- Added: alert notification functionality allows you to receive notifications about the health and important events of your Rocket Pool Smart Node. Check the guides for more details. Thanks to @activescott for this contribution!
- Fixed: approximate RPL reward panel for RPIP-30 changes. Thanks to @cayosenzo for this contribution!
- Fixed: "version is obsolete" issue. This upgrade will automatically remove the obsolete references if your system uses the default install location (
~/.rocketpool
). If you still see warnings, run the following commands, adjusting to your RP directory:
find ~/.rocketpool/override/ -name "*.yml" -exec sed -i '/^version: "3\.7"$/d' {} \;
find ~/.rocketpool/templates/ -name "*.tmpl" -exec sed -i '/^version: "3\.7"$/d' {} \;
v.1.12.1
This is v1.12.1 of the Smart Node.
It is a recommended update for all users as it contains important new features and bug fixes.
Client Updates
- Lighthouse updated to v5.1.2 - includes a block lookup efficiency fix, a fix to stop propagation of slashings for attesters who have already been slashed, a fix to the
--disable-upnp
flag, and a couple of metric fixes; - Teku updated to v24.3.1 - includes improvements to CPU and NETWORK issues. Also introduces Validator Slashing Prevention feature.
- Prysm updated to v5.0.2 - This release has many optimizations, UX improvements, and bug fixes.
- Nimbus updated to v24.3.0 - includes additional beacon API support and resilience to suboptimal network conditions.
- Reth updated to v0.2.0-beta.4 - Improved performance and bug fixes. (Testnet only)
Note that some clients, like Teku, Prysm and Nimbus, started to prioritize locally produced blocks unless MEV-Boost provides a 10% higher value. We are supporting this decision from the client teams and the Smart Node will use these default values.
Smart Node Changes
- Added: alert notification functionality allows you to receive notifications about the health and important events of your Rocket Pool Smart Node. Check the guides for more details. Thanks to @activescott for this contribution!
- Fixed: approximate RPL reward panel for RPIP-30 changes. Thanks to @cayosenzo for this contribution!
- Fixed: "version is obsolete" issue. This upgrade will automatically remove the obsolete references if your system uses the default install location (
~/.rocketpool
). If you still see warnings, run the following commands, adjusting to your RP directory:
find ~/.rocketpool/override/ -name "*.yml" -exec sed -i '/^version: "3\.7"$/d' {} \;
find ~/.rocketpool/templates/ -name "*.tmpl" -exec sed -i '/^version: "3\.7"$/d' {} \;
Client Updates
- Lighthouse updated to v5.1.2 - includes a block lookup efficiency fix, a fix to stop propagation of slashings for attesters who have already been slashed, a fix to the
--disable-upnp
flag, and a couple of metric fixes; - Teku updated to v24.3.1 - includes improvements to CPU and NETWORK issues. Also introduces Validator Slashing Prevention feature.
- Prysm updated to v5.0.2 - This release has many optimizations, UX improvements, and bug fixes.
- Nimbus updated to v24.3.0 - includes additional beacon API support and resilience to suboptimal network conditions.
- Besu updated to v24.4.0 - includes a new RocksDB format. The migration to the new format happens automatically but a downgrade will need an additional command.
- Reth updated to v0.2.0-beta.4 - Improved performance and bug fixes. (Testnet only)
Note that some clients, like Teku, Prysm and Nimbus, started to prioritize locally produced blocks unless MEV-Boost provides a 10% higher value. We are supporting this decision from the client teams and the Smart Node will use these default values.
Smart Node Changes
- Added: alert notification functionality allows you to receive notifications about the health and important events of your Rocket Pool Smart Node. Check the guides for more details. Thanks to @activescott for this contribution!
- Fixed: approximate RPL reward panel for RPIP-30 changes. Thanks to @cayosenzo for this contribution!
- Fixed: "version is obsolete" issue. This upgrade will automatically remove the obsolete references if your system uses the default install location (
~/.rocketpool
). If you still see warnings, run the following commands, adjusting to your RP directory:
find ~/.rocketpool/override/ -name "*.yml" -exec sed -i '/^version: "3\.7"$/d' {} \;
find ~/.rocketpool/templates/ -name "*.tmpl" -exec sed -i '/^version: "3\.7"$/d' {} \;
v1.13.0-dev
This is a development prototype for v1.13.0.
This is actively under development and will change frequently, DO NOT use this unless you were instructed to do so or are willing to break things.
v1.11.9
This is v1.11.9 of the Smartnode.
It is a recommended update for all users as it contains important client updates. It's a mandatory update if you didn't update to v1.11.8 or if you're still running MEV-Boost v1.6.
Client Updates - All the clients are now Dencun-ready!
- MEV-boost updated to v1.7.0 - Deneb ready release for mainnet. Required upgrade for upcoming mainnet dencun fork;
- Grafana updated to v9.5.17 - includes several improvements and bug fixes. Has a known issue related to panel titles not able to be centered (not our fault folks! We waited a long time for a fix);
- Lighthouse updated to v5.1.0 - includes a fix to ensure validator's messages are published on the network. Additionally, out-of-memory crashes are prevented by reducing peak memory usage;
- Lodestar updated to v1.17.0 - includes several bug fixes and new features;
- Besu updated to v24.3.0 - introduces some breaking changes with new parameters;
- Teku updated to v24.3.0 - this is an optional update containing improvements and bug fixes;
- Prysm updated to v5.0.1 - has some nice improvements over the recent v5.0.0;
- Reth updated to v0.2.0-beta.1 - requires a resync (still testnet only).
Smartnode Changes
- Users can now run
rocketpool s prune-eth1
to prune Besu's history. This is an offline process to clean old trie logs and only one execution is needed; - Fixed bug when claiming rewards with no active minipools;
- Add MEV-Boost version in
rocketpool service version
; - Fix: Generate state manager when needed, using the watchtower Archive node if possible;
- New feature: added commands
rocketpool service reset-docker
androcketpool service prune-docker
to clean up docker related resources (old images, containers, networks). Thanks to @activescott for the contribution!
v1.11.9-dev
This is a development prototype for v1.11.9.
This is actively under development and will change frequently, DO NOT use this unless you were instructed to do so or are willing to break things.
v1.11.8
This is v1.11.8 of the Smartnode.
It is a mandatory update for all users as it enables the Deneb/Cancun ("Dencun") Mainnet upgrade, which will occur at March 13, 2024, 01:55:35pm UTC.
Note that with blobs (EIP-4844) being added in Dencun, each node is expected to consume additional disk space, with an expected average of 50 GiB, up to a maximum of 103GiB for blobs.
https://github.com/rocket-pool/smartnode-install/releases/tag/v1.11.8
Client Updates - All the clients are now Dencun-ready and also include improvements and bug fixes!
- Lighthouse updated to v5.0.0 - Improves network parameters and increases the default target peer count to 100;
- Lodestar updated to v1.16.0;
- Geth updated to v1.13.14;
- Besu updated to v24.1.2;
- Teku updated to v24.2.0;
- Nethermind updated to v1.25.4;
- Nimbus updated to v24.2.2 - Has important bug fixes including the one that was causing high CPU load during the backfill process after checkpoint sync. Also fixes the Holesky issue related to the incomplete implementation of EIP-7044;
- Prysm updated to v5.0.0 - brings improvements to the backfill functionality of the beacon node to support backfilling blobs;
- The Reth client v0.1.0-alpha.20 from Paradigm is now available on the Smartnode! The project is still not ready for production and should only be used on the Holesky testnet for the moment. Thanks to Peteris for his work. (note that there's a known issue making Reth fail on some ARM devices)
Smartnode Changes
- Added an Archive mode parameter for Besu, which is disabled by default. When disabled, the
--Xbonsai-trie-log-pruning
parameter will be used, resulting in slower disk space growth; - Fix eth1 metrics for externally managed execution clients;
- Fix removing duplicate pubkeys when fetching validator info;
- Fixed the voluntary exit signatures issue on Holesky;
- Hard-code MaxCollateralFraction for RPIP-30;
- Improved the process to download rewards files with timeouts;
- Make rewardsFileVersion type to encapsulate version range checks;
- Make errors more user friendly.
v1.11.7
This is v1.11.7 of the Smartnode.
It is a recommended update for Mainnet users and a mandatory update for testnet users to be ready for the Holesky Dencun upgrade, which is expected to happen Feb 07th.
Client Updates
- All clients were updated to their latest Dencun-ready versions:
- Geth updated to v1.13.11
- Besu updated to v24.1.1
- Nethermind updated to v1.25.3
- Nimbus updated to v24.2.0 - fixes regressions from v24.1.2
- Lodestar updated to v1.15.0
- Teku updated to 24.1.1
- Prysm updated to v4.2.1 - will cause failures on the Holeski testnet when running on certain older CPUs without AVX support (eg Celeron) after the Deneb fork. This is not an issue for Mainnet.
- Lighthouse updated to v4.6.0 - a medium-priority release with new features, performance improvements and bug fixes for mainnet users. The
--always-prefer-builder-payload
parameter was removed from the BN and moved to the VC asprefer_builder_proposals
.
Smartnode Changes
- Nimbus pruning mode will default to "Prune". This only applies to new users installing the Smartnode. Existing users that want to use this option would need to manually switch via
rocketpool service config
-> ETH2 -> "Pruning mode". It is recommended to checkpoint sync Nimbus after making this change usingrocketpool service resync-eth2
. - The previously deprecated Prater testnet is now completely removed from the Smartnode.
- Nethermind automatic pruning threshold increased from 250 GB to 300 GB.
v1.11.7-dev
This is a development prototype for v1.11.7.
This is actively under development and will change frequently, DO NOT use this unless you were instructed to do so or are willing to break things.
v1.11.6
This is v1.11.6 of the Smartnode.
It is a recommended update for Nethermind and Lodestar operators due to bug fixes and new features related to the updated clients.
Client Updates
- Nethermind updated to v1.25.2 - Fixed eth_syncing invalid behavior. Fixed the mainnet consensus bug introduced on v1.23.0;
- Lodestar updated to v1.14.0 which is recommended for all users;
Smartnode Changes
-
Removed the NethermindPruneStarter dll which broke on the previous version. It was replaced by new Golang code that requests Nethermind to prune without external dependencies;
-
Nethermind will now start full pruning automatically when the volume reaches 250 GB of free space. Manual pruning can still be requested.
-
Added a configuration field to set Nethermind's Full Pruning Memory Budget Size. The Smartnode will select the default according to the node's RAM size.
v1.11.6-dev
This is a development prototype for v1.11.6.
This is actively under development and will change frequently, DO NOT use this unless you were instructed to do so or are willing to break things.