Sui publishes post-mortem after mainnet stall halted transactions for six hours

Sui Basis has revealed an in depth autopsy explaining the reason for the mainnet stall that disrupted transaction processing on 14 January. It confirmed that the community halted as a security measure to stop inconsistent state finalization.

Based on the Basis, the disruption lasted for about six hours. An inner divergence in validator consensus processing induced it. 

Throughout the incident, validators have been unable to certify new checkpoints, resulting in transaction submissions timing out whereas the community prioritized security.

Sui consensus divergence triggered security halt

The Foundation mentioned the incident stemmed from an edge-case bug within the consensus commit logic that affected how conflicting transactions have been dealt with beneath sure garbage-collection circumstances. 

Consequently, totally different validators derived totally different consensus outputs and tried to execute incompatible candidate checkpoints.

When validators detected that greater than one-third of stake was signing a distinct checkpoint digest, checkpoint certification turned unattainable. Validators then halted progress to keep away from finalizing an inconsistent state.

“That is the meant failure mode for this class of concern,” the Basis mentioned, noting that the community is designed to cease safely relatively than threat forks or irreversible inconsistencies.

No forks, rollbacks, or fund losses

Sui careworn that the stall was not brought on by community congestion, transaction quantity, or exterior threats. All through the incident:

• No licensed state forks occurred
• No licensed transactions have been rolled again
• Consumer funds have been by no means in danger
• Community security and consistency ensures have been preserved

Whereas transaction execution halted in the course of the incident window, learn operations continued to serve the final licensed state. This ensured information consistency for customers and functions.

Enhancements deliberate after incident

The Sui Basis mentioned it’s implementing a number of modifications to scale back restoration time within the occasion of comparable points sooner or later. 

Deliberate enhancements embody quicker detection of checkpoint inconsistencies and extra automated operator tooling to wash up divergent inner state. Additionally, expanded consensus-specific testing to breed and validate fixes earlier than deployment.

The Basis added that whereas the interruption was disruptive, it confirmed that Sui’s safety-focused structure behaved as designed.

Remaining Ideas

• Sui’s clarification confirms the mainnet stall was the results of a consensus edge case, with security mechanisms halting the community to keep away from inconsistent finalized state.
• Whereas disruptive, the incident highlights the trade-off between availability and security as high-throughput networks push efficiency limits.