This content is associated with a legacy version of the Replicated product. For the current Replicated product documentation, see docs.replicated.com.
Below is an example of the
swarm section of an application config YAML.
swarm: minimum_node_count: 4 nodes: - role: manager labels: mongo: primary minimum_count: 1 - role: worker labels: mongo: secondary minimum_count: 2
minimum_node_count is specified and there are an insufficient number of nodes in the cluster, the Cluster page of the Replicated Console will render a node count label in orange to indicate that more nodes should be added.
The label will switch to green when the required number of nodes are in the cluster.
Having an insufficient number of nodes will not prevent Replicated from attempting to start the application.
swarm.nodes section is a list of requirements that apply to groups of one or more nodes.
Each group will have a label at the top of the Cluster page that will render in orange or green depending on whether all requirements for the group have been met.
The number of nodes that should have the role and labels specified
A string map.
The Cluster page provides UI for adding and removing labels on individual nodes, equivalent to running
docker node update --label-add and
docker node update --label-rm on the command line.
Use Swarm’s placement constraints and preferences to assign tasks to specific nodes in a cluster. This feature can be used to ensure that stateful services depending on volumes are not rescheduled to another node. When a snapshot is taken of a volume, Replicated stores the role and labels of the node that volume was on. This allows Replicated to restore the volume to the correct node for the service when recovering from a snapshot.
services: mongoreplicas: image: mongo:3.2 ports: - "27017" volumes: - mongodata:/data deploy: replicas: 2 mode: replicated placement: constraints: [node.labels.mongo==secondary] preferences: - spread: node.labels.mongo