Solana remained online after an infrastructure failure temporarily knocked 102 of its 699 staked validators offline, allowing the network to continue producing blocks and processing transactions without a coordinated restart.
Solana Foundation technology executive Jacob Creech said 597 of the network’s 699 staked validators continued voting, while affected validators recovered within about 40 minutes. He confirmed that the incident occurred on August 12 and did not stop Solana from producing blocks or processing transactions.
“Blocks kept producing and transactions kept landing,” Creech said, describing the incident as a test of Solana’s infrastructure resilience. While the blockchain continued operating, data from staking platform Marinade Finance showed that the impact on Solana’s voting stake was considerably larger than the validator count alone suggested.
Staked SOL impact was larger
Analysis from staking platform Marinade Finance, as covered by The Crypto Times, found that 28.83% of staked SOL became delinquent for roughly 33 minutes during the disruption. That brought the network closer to Solana’s 33.34% threshold for transaction finality. Solana can continue producing blocks while a portion of its validators is offline, but more than two-thirds of the network’s stake must participate for transactions to reach finality.
The incident therefore left a relatively narrow margin before finality could have been disrupted. Marinade estimated that around 90 validators were affected by the routing failure, while Creech reported that 102 validators temporarily stopped voting. The difference reflects separate measurements and does not mean all 102 validators were necessarily affected by the same infrastructure failure. The affected validators recovered without requiring Solana to restart its mainnet.
Routing fault disrupts multiple regions
The disruption originated from a networking problem at Teraswitch rather than from a failure in Solana’s blockchain software.
As reported in The Crypto Times preceding report, according to Teraswitch, a malformed default route was advertised from its Miami facility and propagated through a route reflector in Amsterdam. Routers at several European and Asia-Pacific locations then preferred the incorrect route, disrupting connectivity to validators operating at those sites.The disruption affected 12 sites, including locations in London, Amsterdam, Dublin, Frankfurt, Singapore, and Tokyo. North American sites were largely unaffected.
Engineers identified the problematic route within roughly 10 minutes and removed the Miami site from the private backbone. Connectivity gradually returned, with full recovery taking up to about 33 minutes for some affected operators.
Teraswitch has since introduced a global configuration change intended to prevent a similar routing problem from blocking traffic across its infrastructure. The provider said its investigation into the underlying cause remains ongoing.
Solana’s hidden infrastructure risk
The incident highlighted a distinction between decentralization at the validator level and decentralization at the infrastructure level.
Solana has hundreds of independent validators, but those operators can still rely on the same hosting providers, autonomous systems or data centers. If a common provider experiences a major networking failure, multiple independent validators can disappear from the consensus process at the same time.
Marinade found that one autonomous system hosted more than one-quarter of all staked SOL, with about 94% of that stake going offline during the incident. That concentration meant the effect on Solana’s consensus was far larger than the number of affected operators alone would suggest.
Marinade said it plans to review limits on stake concentration across autonomous systems and data centers. It also plans to provide more visibility into whether validators have automatic failover or backup infrastructure.
Different from Solana’s earlier outages
The incident is particularly notable given Solana’s history of network interruptions. In February 2024, Solana’s Mainnet-Beta experienced a major outage after block production halted, prompting core engineers and validators to investigate the issue and prepare a network restart. Validators created snapshots from their local ledger state before restarting the network. The outage lasted several hours before block production resumed.
This time, Solana did not stop producing blocks and did not require a coordinated restart. The failure occurred in the infrastructure used to connect validators rather than in the blockchain’s consensus process.
What comes next
The immediate disruption has been resolved, and there was no indication that user funds were lost because of the outage. The main economic impact fell on affected validators, which missed staking rewards while offline.
The bigger issue is whether Solana can reduce the amount of stake exposed to a single infrastructure provider or network path. Creech said the distribution of validators across independent infrastructure providers helped the network withstand the failure. But the incident also showed that infrastructure diversity can have limits when a large share of stake remains exposed to the same underlying network dependency.
Teraswitch is continuing its investigation and is expected to provide a fuller explanation of the routing failure. Validator operators and staking platforms are also likely to reassess their redundancy and failover arrangements.
Also Read: Alameda’s $352K Solana (SOL) Stake Is Now Worth $15M — Where Are the Funds Going?
