galera_new_cluster ignores the instance name argument and always restarts mariadb.service (regression since 10.6.28)

The documentation says a specific instance can be bootstrapped with "galera_new_cluster mariadb@node1". Since 10.6.28 the script ignores its argument and always runs "systemctl restart mariadb.service". The instance given on the command line is not bootstrapped, and the default mariadb.service is restarted instead. The commit that introduced this changed "systemctl restart ${1:-mariadb}" to "systemctl restart mariadb.service" when the bootstrap env file was moved to the run directory.