PANews June 27 news, Base released a post-mortem report stating that the mainnet experienced two block production interruptions on June 25 and 26, lasting approximately 116 minutes and 20 minutes respectively. The official emphasized that the incident did not affect the security of on-chain assets, and user funds remained available at all times.

The core cause of the failure was a flaw in the sequencer’s block building logic: after a transaction execution failed, the system did not properly clean up the historical journal state, causing abnormal gas calculations when subsequent legitimate transactions were executed, thereby generating invalid state transition blocks and halting block production across the entire L2 network.

During the interruption, the network experienced issues including a complete halt of block production, inability to include transactions on-chain, and mempool congestion. eth_sendRawTransaction requests submitted by users continuously returned errors.

The official subsequently fixed the issue through a patch (PR #3806), restoring block production. However, due to an engine reset race condition during the sequencer cluster restart process, the recovery synchronization was hindered, which also became an indirect cause of the brief outage the next day.

Going forward, Base stated it will focus on strengthening protocol-level fuzzing and stress testing capabilities to detect anomalous transaction paths earlier, while also upgrading monitoring and operations systems, and introducing a more “graceful recovery” mechanism to enhance the network’s rapid recovery capability in similar failures in the future.