Happened to more than half of the 10 nodes we deployed, no commonalities between them, they just hang at some point during eosio_build process. Stopping and restarting the ./start.sh script fixed the issue in every single case.