Skip to content

op-stack v1.7.4

Compare
Choose a tag to compare
@sebastianst sebastianst released this 26 Apr 18:09
· 1468 commits to develop since this release
24a8d3e

⚠️ Strongly recommended maintenance release

🐞 op-node blob reorg bug fix (#10210)

If an L1 block got reorg'd out during blob retrieval, an op-node might get stuck in a loop retrieving a blob that will never exist, requiring a restart. This got fixed by internally signaling the right error types, forcing a derivation pipeline reset in such cases.

✨ op-batcher & op-proposer node sync start (#10116 #10193 #10262 #10273)

op-batcher and op-proposer can now wait for the sequencer to sync to the current L1 head before starting their work.
This fixes an issue where a restart of op-batcher/proposer and op-node at the same time might cause to resend duplicate batches from the last finalized L2 block, because a freshly restarted op-node resyncs from the finalized head, potentially signaling a too early safe head in its sync status.

🏳️ This feature is off by default, so we recommend testing it by using the new batcher and proposer flag --wait-node-sync (or its corresponding env vars).

Enabling this will cause op-batcher and op-proposer to wait for the sequencer's verifier confirmation depth for typically 4 L1 blocks, or ~1 min, at startup.

🏳️ To speed up this process in case that no recent batcher transaction have happened, there's another optional new batcher flag --check-recent-txs-depth that lets the batcher check for recent batcher transactions to determine a potentially earlier sync target. This feature is off by default (0) and should be set to the sequencer's verifier confirmation depth to get enabled.

Partial changelog

op-node

op-batcher & op-proposer

New Contributors (all monorepo)

Full Changelog: v1.7.3...v1.7.4

🚢 Docker Images: