Solana: The downgrade of the mainnet Beta cluster on February 25th was due to the congestion of the main block propagation protocol "Turbine"

Solana issued a statement saying that at about 13:46:16 on February 25th, Beijing time, the reason for the downgrade of the Solana mainnet Beta cluster was the congestion of the main block propagation protocol "Turbine". Most of the block data is transmitted through the extremely slow backup block repair protocol. Core engineers have determined that this issue is caused by a failure in the deduplication logic in the shard forwarding service. Also, the deduplication filter in the retransmission pipeline was not originally designed to prevent loops in the Turbine tree. The deduplication logic is now enhanced to mitigate saturation of this filter in Solana Labs validator client v1.13.7 and v1.14.17. Core engineers are also working with shard forwarding service providers to improve the resiliency and compatibility of their designs. Additionally, a patch has been applied to the Solana Labs validator client that will cause block producers producing large blocks to abort, thereby signaling node operators to check their nodes and bring them back to a healthy state in the cluster. Foresight News previously reported that on February 25, according to CoinDesk, a fork event occurred on the Solana network that affected TPS. After the first attempt failed, the problem evolved from slow transaction processing to almost complete interruption of activities on Solana. Verifiers and Solana The developers at Labs coordinated the second reboot and decided to reboot from socket 179526403 instead of 179526408.