Lifecycle Manager supported capabilities
Lifecycle Manager (LCM) provides powerful abilities for provisioning and configuring DataStax Enterprise (DSE) clusters. The following questions and answers describe the supported capabilities of LCM.
LCM supports creating DSE clusters with up to 300 nodes. Larger clusters cannot be managed using the LCM web interface. The OpsCenter APIs can be used directly to attempt to manage larger clusters, provided sufficient heap memory is available.
OpsCenter supports decommissioning a node. LCM does not currently automate node decommission.
After decommissioning a node, you must delete the corresponding node model in LCM prior to running the next configure or install job. Otherwise, LCM attempts to restore the old topology, with unpredictable results.
OpsCenter supports monitoring multi-instance nodes.
Lifecycle Manager does not currently support managing DSE Multi-Instance nodes (also referred to as dense nodes).
At this time, only the GossipingPropertyFileSnitch (GPFS) is supported for managing or importing DSE clusters in LCM because it is the only snitch that supports all possible topologies, and that can be configured independently on each node.
GPFS is automatically configured according to the rack setting on each node in the cluster topology. For information about migrating to GPFS, see Migrating from the PropertyFileSnitch to the GossipingPropertyFileSnitch.
No. LCM relies on SSH to communicate with target nodes, and requires target nodes to have a full, traditional Linux operating system. LCM works best with IaaS, virtual machines, or bare metal targets.