Cluster Commands

sbctl cluster --help

Cluster Commands

Creates a new cluster.

Created a new control plane cluster with the current node as the primary control plane node.

sbctl cluster create
    --cap-warn=<CAP_WARN>
    --cap-crit=<CAP_CRIT>
    --prov-cap-warn=<PROV_CAP_WARN>
    --prov-cap-crit=<PROV_CAP_CRIT>
    --ifname=<IFNAME>
    --mgmt-ip=<MGMT_IP>
    --tls-secret-name=<TLS_SECRET_NAME>
    --log-del-interval=<LOG_DEL_INTERVAL>
    --metrics-retention-period=<METRICS_RETENTION_PERIOD>
    --contact-point=<CONTACT_POINT>
    --grafana-endpoint=<GRAFANA_ENDPOINT>
    --data-chunks-per-stripe=<DATA_CHUNKS_PER_STRIPE>
    --parity-chunks-per-stripe=<PARITY_CHUNKS_PER_STRIPE>
    --ha-type=<HA_TYPE>
    --is-single-node
    --mode=<MODE>
    --ingress-host-source=<INGRESS_HOST_SOURCE>
    --dns-name=<DNS_NAME>
    --enable-node-affinity
    --fabric=<FABRIC>
    --strict-node-anti-affinity
    --name=<NAME>
    --qpair-count=<QPAIR_COUNT>
    --client-qpair-count=<CLIENT_QPAIR_COUNT>
    --client-data-nic=<CLIENT_DATA_NIC>
    --use-backup=<USE_BACKUP>
    --nvmf-base-port=<NVMF_BASE_PORT>
    --rpc-base-port=<RPC_BASE_PORT>
    --snode-api-port=<SNODE_API_PORT>

Parameter	Description	Data Type	Required	Default
--cap-warn	The capacity warning level in percent. Default: `89`.	integer	False	89
--cap-crit	The capacity critical level in percent. Default: `99`.	integer	False	99
--prov-cap-warn	The capacity warning level in percent. Default: `250`.	integer	False	250
--prov-cap-crit	The capacity critical level in percent. Default: `500`.	integer	False	500
--ifname	Management interface name, e.g. eth0.	string	False	-
--mgmt-ip	Management IP address to use for the node (e.g., 192.168.1.10).	string	False	-
--tls-secret-name	Name of the Kubernetes TLS Secret to be used by the Ingress for HTTPS termination (e.g., my-tls-secret).	string	False	-
--log-del-interval	The logging retention policy. Default: `3d`.	string	False	3d
--metrics-retention-period	Retention period for I/O statistics (Prometheus). Default: `7d`.	string	False	7d
--contact-point	The email or slack webhook url to be used for alerting.	string	False
--grafana-endpoint	The endpoint url for Grafana.	string	False
--data-chunks-per-stripe	The erasure coding schema parameter k (distributed raid). Default: `1`.	integer	False	1
--parity-chunks-per-stripe	The erasure coding schema parameter n (distributed raid). Default: `1`.	integer	False	1
--ha-type	Logical volume HA type (single, ha), default is cluster ha type. Default: `ha`. Available Options: - single - ha	string	False	ha
--is-single-node	For single-node clusters only. Default: `false`.	marker	False	False
--mode	The environment to deploy management services. Default: `docker`. Available Options: - docker - kubernetes	string	False	docker
--ingress-host-source	Ingress host source: 'hostip' for node IP, 'loadbalancer' for external LB, or 'dns' for custom domain. Default: `hostip`. Available Options: - hostip - loadbalancer - dns	string	False	hostip
--dns-name	Fully qualified DNS name to use as the Ingress host (required if --ingress-host-source=dns).	string	False
--enable-node-affinity	Enable node affinity for storage nodes.	marker	False	-
--fabric	The NVMe fabric to use (specify: `tcp`, `rdma`, `tcp,rdma`). Default: `tcp`. Available Options: - tcp - rdma - tcp,rdma	string	False	tcp
--strict-node-anti-affinity	Enable strict node anti affinity for storage nodes. Never more than one chunk is placed on a node. This requires a minimum of data-chunks-in-stripe + parity-chunks-in-stripe + 1 nodes in the cluster.	marker	False	-
--name, -n	Assigns a name to the newly created cluster.	string	False	-
--qpair-count	Increase for clusters with few but very large logical volumes or decrease for clusters with a large number of very small logical volumes.	range(0..128)	False	32
--client-qpair-count	Increase for clusters with few but very large logical volumes or decrease for clusters with a large number of very small logical volumes.	range(0..128)	False	3
--client-data-nic	Network interface name from client to use for logical volume connection.	string	False	-
--use-backup	The path to JSON file with S3/MinIO backup configuration.	string	False	-
--nvmf-base-port	Base port for all NVMe-oF listeners (lvol, hublvol, device). Default: `4420`.	integer	False	4420
--rpc-base-port	The base port for SPDK JSON-RPC. Default: `8080`.	integer	False	8080
--snode-api-port	The SNodeAPI/firewall port (one per host IP). Default: `50001`.	integer	False	50001

Adds a new cluster.

sbctl cluster add
    --cap-warn=<CAP_WARN>
    --cap-crit=<CAP_CRIT>
    --prov-cap-warn=<PROV_CAP_WARN>
    --prov-cap-crit=<PROV_CAP_CRIT>
    --data-chunks-per-stripe=<DATA_CHUNKS_PER_STRIPE>
    --parity-chunks-per-stripe=<PARITY_CHUNKS_PER_STRIPE>
    --ha-type=<HA_TYPE>
    --enable-node-affinity
    --fabric=<FABRIC>
    --is-single-node
    --qpair-count=<QPAIR_COUNT>
    --client-qpair-count=<CLIENT_QPAIR_COUNT>
    --strict-node-anti-affinity
    --name=<NAME>
    --client-data-nic=<CLIENT_DATA_NIC>
    --use-backup=<USE_BACKUP>
    --nvmf-base-port=<NVMF_BASE_PORT>
    --rpc-base-port=<RPC_BASE_PORT>
    --snode-api-port=<SNODE_API_PORT>

Parameter	Description	Data Type	Required	Default
--cap-warn	The capacity warning level in percent. Default: `89`.	integer	False	89
--cap-crit	The capacity critical level in percent. Default: `99`.	integer	False	99
--prov-cap-warn	The capacity warning level in percent. Default: `250`.	integer	False	250
--prov-cap-crit	The capacity critical level in percent. Default: `500`.	integer	False	500
--data-chunks-per-stripe	The erasure coding schema parameter k (distributed raid). Default: `1`.	integer	False	1
--parity-chunks-per-stripe	The erasure coding schema parameter n (distributed raid). Default: `1`.	integer	False	1
--ha-type	Logical volume HA type (single, ha), default is cluster single type. Default: `ha`. Available Options: - single - ha	string	False	ha
--enable-node-affinity	Enables node affinity for storage nodes.	marker	False	-
--fabric	Fabric: tcp, rdma or both (specify: tcp, rdma). Default: `tcp`. Available Options: - tcp - rdma - tcp,rdma	string	False	tcp
--is-single-node	For single-node clusters only. Default: `false`.	marker	False	False
--qpair-count	Increase for clusters with few but very large logical volumes or decrease for clusters with a large number of very small logical volumes.	range(0..128)	False	32
--client-qpair-count	Increase for clusters with few but very large logical volumes or decrease for clusters with a large number of very small logical volumes.	range(0..128)	False	3
--strict-node-anti-affinity	Enable strict node anti affinity for storage nodes. Never more than one chunk is placed on a node. This requires a minimum of data-chunks-in-stripe + parity-chunks-in-stripe + 1 nodes in the cluster."	marker	False	-
--name, -n	Assigns a name to the newly created cluster.	string	False	-
--client-data-nic	Network interface name from client to use for logical volume connection.	string	False	-
--use-backup	The path to JSON file with S3/MinIO backup configuration.	string	False	-
--nvmf-base-port	Base port for all NVMe-oF listeners (lvol, hublvol, device). Default: `4420`.	integer	False	4420
--rpc-base-port	The base port for SPDK JSON-RPC. Default: `8080`.	integer	False	8080
--snode-api-port	The SNodeAPI/firewall port (one per host IP). Default: `50001`.	integer	False	50001

Activates a cluster.

Once a cluster has sufficient nodes added, it needs to be activated. Can also be used to re-activate a suspended cluster.

sbctl cluster activate
    <CLUSTER_ID>
    --force
    --force-lvstore-create

Argument	Description	Data Type	Required
CLUSTER_ID	The cluster id.	string	True

Parameter	Description	Data Type	Required	Default
--force	Force recreate distr and lv stores.	marker	False	-
--force-lvstore-create	Force recreate lv stores.	marker	False	-

Shows the cluster list.

sbctl cluster list
    --json

Parameter	Description	Data Type	Required	Default
--json	Print json output.	marker	False	-

Shows a cluster's status.

sbctl cluster status
    <CLUSTER_ID>

Argument	Description	Data Type	Required
CLUSTER_ID	The cluster id.	string	True

Create lvstore on newly added nodes to the cluster.

sbctl cluster complete-expand
    <CLUSTER_ID>

Argument	Description	Data Type	Required
CLUSTER_ID	The cluster id.	string	True

Shows a cluster's statistics.

sbctl cluster show
    <CLUSTER_ID>

Argument	Description	Data Type	Required
CLUSTER_ID	The cluster id.	string	True

Gets a cluster's information.

sbctl cluster get
    <CLUSTER_ID>

Argument	Description	Data Type	Required
CLUSTER_ID	The cluster id.	string	True

Gets a cluster's capacity.

sbctl cluster get-capacity
    <CLUSTER_ID>
    --json
    --history=<HISTORY>

Argument	Description	Data Type	Required
CLUSTER_ID	The cluster id.	string	True

Parameter	Description	Data Type	Required	Default
--json	Print json output.	marker	False	-
--history	(XXdYYh), list history records (one for every 15 minutes) for XX days and YY hours (up to 10 days in total).	string	False	-

Gets a cluster's I/O statistics.

sbctl cluster get-io-stats
    <CLUSTER_ID>
    --records=<RECORDS>
    --history=<HISTORY>

Argument	Description	Data Type	Required
CLUSTER_ID	The cluster id.	string	True

Parameter	Description	Data Type	Required	Default
--records	The number of records. Default: `20`.	integer	False	20
--history	(XXdYYh), list history records (one for every 15 minutes) for XX days and YY hours (up to 10 days in total).	string	False	-

Returns a cluster's status logs.

sbctl cluster get-logs
    <CLUSTER_ID>
    --json
    --limit=<LIMIT>

Argument	Description	Data Type	Required
CLUSTER_ID	The cluster id.	string	True

Parameter	Description	Data Type	Required	Default
--json	Return JSON formatted logs.	marker	False	-
--limit	Show last number of logs, default 50. Default: `50`.	integer	False	50

Gets a cluster's secret.

sbctl cluster get-secret
    <CLUSTER_ID>

Argument	Description	Data Type	Required
CLUSTER_ID	The cluster id.	string	True

Updates a cluster's secret.

sbctl cluster update-secret
    <CLUSTER_ID>
    <SECRET>

Argument	Description	Data Type	Required
CLUSTER_ID	The cluster id.	string	True
SECRET	The new 20 characters password.	string	True

Updates a cluster's fabric.

sbctl cluster update-fabric
    <CLUSTER_ID>
    <FABRIC>

Argument	Description	Data Type	Required
CLUSTER_ID	The cluster id.	string	True
FABRIC	Fabric: tcp, rdma or both (specify: tcp, rdma). Default: `tcp`.	string	True

Checks a cluster's health.

sbctl cluster check
    <CLUSTER_ID>

Argument	Description	Data Type	Required
CLUSTER_ID	The cluster id.	string	True

Updates a cluster to new version.

Updates a the control plane to a new version. To update the storage nodes, they have to be shutdown and restarted. This can be done in a rolling manner. Attention: verify that an upgrade path is available and has been tested!"

sbctl cluster update
    <CLUSTER_ID>
    --cp-only=<CP_ONLY>
    --spdk-image=<SPDK_IMAGE>
    --mgmt-image=<MGMT_IMAGE>

Argument	Description	Data Type	Required
CLUSTER_ID	The cluster id.	string	True

Parameter	Description	Data Type	Required	Default
--cp-only	Update the control plane only. Default: `false`.	boolean	False	False
--spdk-image	Restart the storage nodes using the provided image.	string	False	-
--mgmt-image	Restart the management services using the provided image.	string	False	-

Lists tasks of a cluster.

sbctl cluster list-tasks
    <CLUSTER_ID>
    --limit=<LIMIT>

Argument	Description	Data Type	Required
CLUSTER_ID	The cluster id.	string	True

Parameter	Description	Data Type	Required	Default
--limit	Show last number of tasks, default 50. Default: `50`.	integer	False	50

Cancels task by task id.

sbctl cluster cancel-task
    <TASK_ID>

Argument	Description	Data Type	Required
TASK_ID	The cluster task id.	string	True

Get rebalancing subtasks list.

sbctl cluster get-subtasks
    <TASK_ID>

Argument	Description	Data Type	Required
TASK_ID	The cluster task id.	string	True

Deletes a cluster.

This is only possible, if no storage nodes and pools are attached to the cluster

sbctl cluster delete
    <CLUSTER_ID>

Argument	Description	Data Type	Required
CLUSTER_ID	The cluster id.	string	True

Assigns or changes a name to a cluster

sbctl cluster change-name
    <CLUSTER_ID>
    <NAME>

Argument	Description	Data Type	Required
CLUSTER_ID	The cluster id.	string	True
NAME	The new cluster name.	string	True

Assigns the snapshot replication target cluster

sbctl cluster add-replication
    <CLUSTER_ID>
    <TARGET_CLUSTER_ID>
    --timeout=<TIMEOUT>
    --target-pool=<TARGET_POOL>

Argument	Description	Data Type	Required
CLUSTER_ID	Cluster id	string	True
TARGET_CLUSTER_ID	Target Cluster id	string	True

Parameter	Description	Data Type	Required	Default
--timeout	Snapshot replication network timeout	integer	False	3600
--target-pool	Target cluster pool ID or name	string	False	-