Optimism shipped five OP Stack releases today. Read the op-reth diff, not the changelog.
Op-reth v2.2.3 adds a new historical proofs storage format. Switch to it without reinitializing your data first, and your fault proofs break.
CryptoVibe Desk · optimism · op-stack · fault-proofs
- →Optimism published five OP Stack releases today, including op-reth v2.2.3, kona-node v1.5.1, and op-conductor v0.9.4.
- →Op-reth v2.2.3 introduces storage format v2 with a required reinit step, while kona-node v1.5.1 patches key fault-proof execution paths including preimage hash verification.
- →Watch for proof disputes on live OP Stack chains that fail to resolve, and for operators who quietly miss op-reth's reinit requirement on upgrade.
- fault proof → A system that lets anyone challenge a sequencer's transaction output on-chain by re-running the disputed block in a verifiable sandboxed environment.
- preimage → Raw data that produces a specific hash committed on-chain; in Optimism's proof system, nodes must verify this match before trusting the data in a proof.
- FPVM → The Fault Proof Virtual Machine, the sandboxed environment where Optimism re-runs disputed transactions to verify them without relying on the sequencer.
- Glamsterdam → An upcoming Ethereum execution upgrade; OP Stack clients must update their internal gas accounting rules before it activates to keep proofs valid.
Optimism released five OP Stack components today. Most are minor. Two require attention.
Op-reth v2.2.3, the Rust execution client, adds historical proofs storage format v2. Enabling it takes one flag: `--proofs-history.storage-version=v2`. The release notes are clear: proofs data must be reinitialized when switching formats. If you enable the flag on an existing node without reinitializing, your historical proofs become inconsistent. That's the catch for any operator whose node feeds into Optimism's fault-proof pipeline.
Kona-node v1.5.1 is tagged recommended. The fixes cover the fault-proof execution path: preimage hash verification, invalid frame-byte rejection, FPVM precompile oracle gas for Glamsterdam, and state-trie handling.
The preimage verification fix is the one worth understanding. In Optimism's proof system, a prover commits a hash on-chain and later supplies the preimage. If the node accepted a mismatched preimage, the proof could fail silently. This patch closes that path.
Op-conductor v0.9.4 manages sequencer failover in multi-sequencer setups. The release notes don't describe specific changes beyond the version bump. If you operate a sequencer cluster, treat this as a required maintenance upgrade.
The two remaining releases, op-batcher v1.16.9 and op-node v1.18.1, are both optional. One has logging changes, one has metrics changes. Skip them unless you need the observability fix.
The bigger picture is what Optimism's multi-prover roadmap looks like from inside the diffs. Independent clients need to agree on data formats, proof encodings, and gas accounting. That's required before the fault-proof system can run reliably across different execution environments.
Storage versioning in op-reth, hash verification in kona-node, gas accounting for Glamsterdam. It doesn't look like a roadmap milestone from the outside. It is one.
If you're running op-reth today, the v2 storage format is opt-in for now. The tradeoff is a one-time migration for a cleaner layout going forward. Wait and you're reinitializing under pressure instead of on your own schedule.
Leaving storage format v2 as a silent opt-in without a startup check is Optimism's oversight: operators who miss the reinit step won't know until their proofs are already inconsistent.
A fault-proof dispute that fails to resolve on any live OP Stack chain within the next 90 days, especially on nodes running pre-v1.5.1 kona-node, would confirm the preimage verification fix addressed a real production gap rather than a theoretical one.
Primary links and supporting reads used by the desk for this story.
Forward this.
