Between blocks 8,300,569 and 8,300,570, there was a network failure that caused approximately 50% of Cardano nodes to shut down and restart. This is written by a development team from Input Output Global (IOG).
The failure affected repeaters and block producers. The developers also clarified that, apparently, the edge nodes were not affected.
The cause of the failure was a wandering problem that led to one of two reactions: some nodes disconnected from peers (peer), others rebooted.
IOG engineers noted that they had foreseen the occurrence of such errors in the design.
“The systems behaved as expected,” they added.
Block production is known to have been “minorly” affected, but part of the network was out of sync until the restart of most of the affected nodes.
“The impact was subtle, similar to delays that also occur during normal operations, and are often observed at the edge of epochs. Most nodes recovered automatically, depending on the type of staking pool operator software deployed,” IOG said.