This is v1.7.4 of the Smartnode stack. This is another low priority update for most users that includes a few small fixes and handy new features. It's a medium priority update for Nethermind users, since it fixes pruning and the new client comes with a lot of improvements.
Client Updates
-
Updated Nethermind to v1.16.1.
- ⚠️ NOTE: If you are running both Nethermind and Prysm together, you will need to add
--JsonRpc.MaxBatchSize=10000to Nethermind's "Additional Flags" field in therocketpool service configTUI until Prysm releases a fix.
- ⚠️ NOTE: If you are running both Nethermind and Prysm together, you will need to add
-
Updated Nimbus to v23.1.0.
Smartnode Changes
-
This introduces Rewards Tree Spec v4, which will go live on February 16th for Interval 6.
- Minipools that were not active on the Beacon chain before the interval checkpoint, and minipools that exited the Beacon chain before the interval checkpoint, are no longer eligible for RPL rewards.
- Oracle DAO members must update to this version (or higher) before February 16th to guarantee that your node has consensus with the others.
-
Reduced Nethermind's cache size by half on systems with less than 32 GB of RAM to help mitigate out-of-memory errors.
-
Fixed metrics reporting for Nethermind, so Grafana can track it again.
-
Removed the Sync Peers Report from Nethermind's logs again so it stops flooding the logs.
-
Nimbus now supports pruning! In the
rocketpool service configTUI, there is a new setting for Nimbus called "Pruning Mode". By default this is set to "Archive" which will store all of the Beacon Chain data, but you can change it to "Pruned" which will discard everything older than 5 months to save on disk space.- ⚠️ NOTE: if you already have a synced Nimbus client and want to switch from Archive to Pruned, it will take a very long time to finish pruning and Nimbus will be offline during this process. We recommend switching from Archive to Pruned, then immediately running
rocketpool service resync-eth2to delete the chain data and do a checkpoint sync to redownload what it needs. Make sure you have a checkpoint sync provider assigned in the "Checkpoint Sync URL" box!
- ⚠️ NOTE: if you already have a synced Nimbus client and want to switch from Archive to Pruned, it will take a very long time to finish pruning and Nimbus will be offline during this process. We recommend switching from Archive to Pruned, then immediately running
-
The daemon's
sign-messagefeature (used by Rocketarb and the rescue node) no longer require your node to be registered with the network or your clients to be synced in order to sign custom messages with your node wallet. -
rocketpool service export-eth1-datanow has a--dirtyflag that lets you export the chain data while your Execution Client is still running. This allows you to do two passes: a--dirtyone followed by a normal one; since the latter will only copy files that have changed since the former, it will be much faster which means less Execution Client downtime.- Thanks to @0xfornax for implementing both of these.
-
Docker-compose will no longer spam the console with "Downloading" and "Extracting" messages after running
rocketpool service start.
Misc Changes
- Updated Go to v1.19.5.