Given a mapped_data extraconfig like this one:
"extraconfig": {
"mapped_data": {
"rabbitmq_config_variables": {
"queue_master_locator": "<<\"min-masters\">>",
"loopback_users": [],
"vm_memory_high_watermark": 0.5,
"rates_mode": "detailed",
"stats_event_max_backlog": 250,
"tcp_listen_options": "[binary, {packet, raw}, {reuseaddr, true}, {backlog, 128}, {nodelay, true}, {exit_on_close, false}, {keepalive, true}]",
"disk_free_limit": 2000000000,
"cluster_partition_handling": "pause_minority",
"collect_statistics_interval": 30000
},
The 40-hiera-datafiles script will convert the large integer 'disk_free_limit' like this:
rabbitmq_config_variables: {
"collect_statistics_interval": 30000,
"queue_master_locator": "<<\"min-masters\">>",
"loopback_users": [],
"vm_memory_high_watermark": 0.5,
"rates_mode": "detailed",
"stats_event_max_backlog": 250,
"tcp_listen_options": "[binary, {packet, raw}, {reuseaddr, true}, {backlog, 128}, {nodelay, true}, {exit_on_close, false}, {keepalive, true}]",
"disk_free_limit": 2e+09,
"cluster_partition_handling": "pause_minority"
}
Which will break the rabbitmq service as it cannot parse '2e+09'. This was first observed on newton
Quoting the large number fixes it