Solana stopped producing blocks as consensus stalled

Solana’s Mainnet Beta cluster stopped producing blocks at approximately 20:30 UTC on April 30, 2022 after its validators lost consensus. The interruption prevented the network from confirming new state and sent operators into a coordinated recovery process that continued beyond the end of April 30 UTC.

The halt mattered because Solana was positioning high throughput and low transaction costs as distinguishing features for decentralized finance, payments and non-fungible-token applications. When consensus stopped, those applications could not obtain newly finalized transactions from the base network, regardless of whether their own software remained online.

A validator-coordinated recovery

The operational record preserved in the “Solana Mainnet-Beta Cluster Restart 30 April 2022” runbook identified slot 131,973,970 as the highest optimistically confirmed slot. It instructed validators to build or obtain a snapshot tied to that slot, verify an expected bank hash and shred version, and wait for 80% of activated stake before resuming validation.

Those instructions show that recovery was not equivalent to restarting one company server. Operators needed a common view of the last acceptable chain state and enough participating stake to resume block production without splitting the network again. The runbook also made filtering transactions associated with the Candy Machine minting program optional, warning validators to remove any temporary filter after stabilization.

The document’s currently accessible version includes later completion notices and may have been updated during and after the incident. It nevertheless preserves the specific slot, bank hash and validator procedure used for the recovery. Coinburn therefore treats it as a primary operational record, not as an immutable timestamped transcript of every decision made on April 30.

What was known on April 30

The verified event-day claim is narrow: block production ceased because consensus stalled, and validators began coordinating a cluster restart. The complete technical cause had not been established publicly before April 30 ended in UTC.

That uncertainty is important. Heavy traffic occurring near a failure does not, by itself, prove whether the decisive fault was malicious denial of service, inadequate traffic controls, validator resource exhaustion, consensus behavior or some combination. Contemporary participants considered temporarily filtering Candy Machine traffic, but that operational response was not yet a complete root-cause analysis.

The halt also should not be described merely as slower processing. A congested chain can continue producing blocks while transactions wait or fail. Solana’s April 30 incident crossed a more serious boundary: the validator cluster ceased advancing its ledger and required coordinated intervention.

No SOL price, return or trading-volume claim is included here. Crypto assets trade continuously across venues, and a defensible market-impact calculation would require a named SOL instrument, venue, currency pair, UTC window and comparison benchmark. The surviving operational records establish the network failure more directly than they establish any causal market move.

Later context: the May 1 restart and May 2 diagnosis

Solana Foundation’s May 2 incident report said block production resumed at 03:30 UTC on May 1, making the interruption approximately seven hours. The report attributed the initiating load to bots competing in a fixed-price NFT mint through Candy Machine. It measured approximately six million inbound transactions per second and more than 100 gigabits per second of traffic at individual nodes.

According to that later diagnosis, validators exhausted memory and crashed as insufficient voting prevented finalization and abandoned-fork cleanup. The accumulating forks exceeded the validators’ capacity and made manual intervention necessary. The Foundation said it found no evidence of a deliberate denial-of-service attack.

Those findings clarify the April 30 event but were not available in completed form when the halt began. The Foundation subsequently outlined QUIC-based transaction transport, stake-weighted quality of service and fee-based execution priority as mitigations under development. On April 30, those were not deployed guarantees, and the restart did not by itself establish that Solana’s congestion risks had been resolved.