Administer database clusters
Mission Control clusters are objects describing the topology and settings of Hyper-Converged Database (HCD), DataStax Enterprise (DSE), and Apache Cassandra® database instances.
Mission Control creates at least one project to contain a logical group of database clusters.
Database cluster management encompasses the entire lifecycle of cluster components. This includes creating datacenter-level resources across Kubernetes cluster boundaries, orchestrating automation steps at appropriate intervals given a specific cluster state, and remediating issues as they arise. Given the distributed nature of Mission Control, there are Kubernetes operators local to each Kubernetes cluster which respond to events within the local cluster. This enables continuous operations and remediation even during periods of network partitioning.
Use Mission Control for database cluster tasks such as configuration, execution, and monitoring status during a cluster lifecycle or a:
-
datacenter lifecycle addressing Add a datacenter to an existing DataStax Enterprise (DSE) cluster or Terminate a datacenter within an existing database cluster.
-
node lifecycle addressing Add nodes to a cluster, Replace a node, Remove nodes from a cluster, or Use per-node configurations.
Interactions with Mission Control are broadly categorized as declarative or imperative. Use declarative APIs to define a desired end-state, and then Mission Control constantly takes steps to achieve that declared goal. For example, you would define a DataStax Enterprise (DSE) or Hyper-Converged Database (HCD) cluster or datacenter through declarative APIs because these are long-lived resources. In comparison, imperative APIs describe a singular operation or task to be run to completion with an associated status. Imperative APIs within Mission Control include operational tasks on a specified cluster, such as performing a rolling restart.
Perform declarative or imperative operations at either the cluster or the datacenter level.
- Cleanup
-
Run the DSE
nodetool cleanup
command for specific keyspaces on all nodes in a particular datacenter (DC). - Rebuild
-
Run the DSE
nodetool rebuild
command across a datacenter to populate data by streaming from another (source) datacenter (DC). - Replace a DSE node
-
Replaces an existing node with a new, empty pod owning the same token ranges.
- Rolling Restart
-
Restart all nodes in a cluster in a rolling fashion while applications continue to run with zero downtime.
- Backup data
-
Backs up Apache Cassandra® and DataStax Enterprise (DSE) data in Kubernetes.
- Restore a data backup
-
Restores Apache Cassandra® and DataStax Enterprise (DSE) data in Kubernetes.
- Upgrade SSTables
-
Upgrade all SSTables to newer formats where required.
Pre-configure a cluster manifest
DataStax recommends storing your cluster manifest in a Git repository to make the cluster provisioning process traceable and repeatable.
Use your configured cluster manifest to [Create a new cluster] using either the UI or a CLI command.
The manifest exposes a partial subset of the MissionControlCluster
resource definition that you can change by editing the cluster manifest.
Access the Mission Control User Interface (UI)
Mission Control provides a UI through the IP address of any node using port 30880
on the Control-Plane
cluster.
For example, issue https://10.0.0.3:30880
from a web browser, where 10.0.0.3
is a qualifying nodes' IP address.
Prerequisites
Because Mission Control organizes clusters by projects, an existing project is required. See create project.