op-node v1.19.8

Overview

Important

This is a recommended release for sequencer operators, particularly those running with --l2.follow.source. It contains a block-building performance fix for sequencers and stricter validation of hardfork activation times in rollup configs.

Operators running nodes on Sepolia OP Stack chains must upgrade to op-node v1.19.5+ prior to the Glamsterdam hardfork on October 6th.

Other changes

• Sequencers now fetch each new L1 head's block and receipts as soon as it is observed, instead of on demand when it becomes the L1 origin. This removes the slow block builds that recurred at L1-origin transitions on sequencers running with --l2.follow.source, where no local derivation had already fetched that data. A failed prefetch logs Failed to prefetch L1 receipts at warn level and the normal fetch still runs (#22863)
• op-node now refuses to start with a rollup config that activates Jovian or any later hardfork at the same timestamp as the fork before it, unless that timestamp is at or before genesis. Every fork is also now checked for activation order, so a config that schedules Jovian, Karst or Lagoon without its predecessor, or before it, is rejected too. No chain in the embedded superchain registry is affected; a custom rollup config that violates the rule must be corrected before upgrading (#22838)

Full Changelog: op-node/v1.19.7...op-node/v1.19.8

🚢 Docker Image:

• https://us-docker.pkg.dev/oplabs-tools-artifacts/images/op-node:v1.19.8