Contact UsGet started
© 2025 Direct Cursus Technology L.L.C.

yc container node-group create

Written by
Updated at December 2, 2024

Create a node group.

Command Usage

Syntax:

yc container node-group create <NODE-GROUP-NAME> [Flags...] [Global Flags...]

Flags

Flag Description
--name string
Name of the node group.
--description string
New description for the node group.
--labels key=value[,key=value...]
List of node group labels as key-value pairs.
--cluster-id string
ID of the Kubernetes cluster.
--cluster-name string
Name of the Kubernetes cluster.
--platform string
Specifies platform for instances in the node group.
--metadata key=value[,key=value...]
Metadata for instances in the node group.
--metadata-from-file key=value[,key=value...]
Same as --metadata except that the value for the entry will be read from a local file.
--template-labels key=value[,key=value...]
List of platform (not Kubernetes) labels for Compute Cloud instances in the node group as key-value pairs.
No more than 32 per resource. The maximum string length in characters for each value is 128. The string length in characters for each key must be 1-63. Each key must match the regular expression '[a-z][-_./\@0-9a-z]*'.
--template-labels-from-file key=value[,key=value...]
Same as --template-labels except that the list of key-value pairs for the entry will be read from a local file.
--public-ip Set when node group needs external public IP.
--memory byteSize
Specifies how much memory instances should have.
--cores int
Specifies how many CPU cores instances should have.
--core-fraction int
Specifies baseline performance for a core in percent.
--gpus int
Specifies how many GPUs instances should have.
--preemptible Specifies if created nodes will be preemptible.
--gpu-cluster-id string
Specifies if created nodes will be part of the GPU cluster.
--gpu-environment string
Specifies GPU environment for created nodes. Values: 'runc-drivers-cuda', 'runc'
--disk-type string
Specifies boot disk type for instances, valid values: network-ssd|network-hdd
--disk-size byteSize
Specifies boot disk size for instances.
--fixed-size int
Specifies fixed scale policy size.
--auto-scale PROPERTY=VALUE[,PROPERTY=VALUE...]
Auto scale policy for Node Group.

Possible property names:
  • min: Minimum number of nodes in the node group.
  • max: Maximum number of nodes in the node group.
  • initial: Initial number of nodes in the node group.
--location PROPERTY=VALUE[,PROPERTY=VALUE...]
List of node group locations.

Possible property names:
  • subnet-id: Subnet id.
  • subnet-name: Subnet name.
  • zone: Zone of the subnet.
--version string
Node Kubernetes components version. Default depends on Cluster release channel. See available in 'yc managed-kubernetes list-versions'.
--allowed-unsafe-sysctls value[,value]
List of strings containing allowed unsafe sysctls for use on nodes.
--node-labels key=value[,key=value...]
Kubernetes labels that are assigned to the nodes of the node group.
--node-taints value[,value]
Kubernetes taints that are applied to the nodes of the node group.
--auto-upgrade Sets auto upgrade maintenance policy. 'true' by default.
Policy defines if Node Group can be upgraded to newer version revision in specified maintenance window.
Auto upgrade is 'graceful': new nodes are created, while old nodes drained by one. Please setup Pod Disruption Budget, for graceful pods drain.
For details about PDB see https://kubernetes.io/docs/tasks/run-application/configure-pdb/
New revision contains can contain bug and security fixes, so it's recommended to leave this policy enabled. Node Group never upgraded to another major version automatically, even with this policy set to true.
Examples: '--auto-upgrade=false', '--auto-upgrade'
--auto-repair Sets auto repair maintenance policy. 'true' by default.
Policy defines if nodes seems to be broken can be recreated automatically.
Examples: '--auto-repair=false', '--auto-repair'
--anytime-maintenance-window Allow maintenance anytime. This is default maintenance window.
--daily-maintenance-window PROPERTY=VALUE[,PROPERTY=VALUE...]
Allow maintenance everyday specified days of week in the specified time window.

Flag can be used multiple times, to define different time windows for different days of week.

Examples:

--daily-maintenance-window 'start=22:00,duration=10h

--daily-maintenance-window 'start=03:00,duration=8h30m

Possible property names:
  • start: Time of day in UTC time zone, HH:MM 24-hour clock format, since when maintenance operations are allowed. Example: '02:30'
  • duration: Maintenance window duration in interval [1h; 24h]. Example: '12h30m'
--weekly-maintenance-window PROPERTY=VALUE[,PROPERTY=VALUE...]
Allow maintenance on specified days of week in the specified time window.

Flag can be used multiple times, to define different time windows for different days of week.

Examples:

--weekly-maintenance-window 'days=[monday,tuesday],start=22:00,duration=10h

--weekly-maintenance-window 'days=weekend,start=03:00,duration=8h30m

Possible property names:
  • days: List of days of the week on which maintenance is allowed in specified time window. Allowed values: weekdays, weekend, monday, tuesday, wednesday, thursday, friday, saturday, sunday, mon, tue, wed, thu, fri, sat, sun, all. Examples: 'weekend', '[weekdays,saturday]', '[monday,wednesday,friday]'
  • start: Time of day in UTC time zone, HH:MM 24-hour clock format, since when maintenance operations are allowed. Example: '02:30'
  • duration: Maintenance window duration in interval [1h; 24h]. Example: '12h30m'
--max-expansion int
Number of extra nodes to be created on each upgrade of the node group.
Must be used in conjunction with --max-unavailable flag.
--max-unavailable int
Number of nodes that can be unavailable at the same time on each upgrade of the node group.
Must be used in conjunction with --max-expansion flag.
--network-interface PROPERTY=VALUE[,PROPERTY=VALUE...]
Adds a network interface to the node group.

Possible property names:
  • subnets: Specifies the subnets for the interface, subnets=[id1,id2],subnets=id3It is expected, that there is one subnet for each node group location zone.Can use either names or subnet ids, or both.
  • ipv4-address: Use 'auto' to assigns internal IPv4 address to the interface.Use 'nat' to use One-to-One NAT on the interface.
  • ipv6-address: Use 'auto' to assigns internal IPv6 address to the interface.Use 'nat' to use One-to-One NAT on the interface.
  • security-group-ids: Security groups for the network interface, security-group-ids=[id1,id2],security-group-ids=id3
--placement-group string
Placement policy group for the node group. Can be either name or ID.
--network-acceleration-type string
Type of a network acceleration for nodes. Values: 'standard', 'software-accelerated'
--container-runtime string
Type of a container runtime for nodes. Values: 'docker', 'containerd'
--node-name string
Name of the node.
In order to be unique it must contain at least one of the instance unique placeholders like {instance_group.id}, {instance.short_id}, {instance.index}.
See Compute service Instance-group metadata doc for full list.
Example: --node-name=prod-node-
--container-network-settings PROPERTY=VALUE[,PROPERTY=VALUE...]
Container network settings for nodes.

Possible property names:
  • pod-mtu: MTU for pods. Value must be in interval 1000 - 8910 if set. Can't be used on clusters with Cilium or Calico.
--async Display information about the operation in progress, without waiting for the operation to complete.

Global Flags

Flag Description
--profile string
Set the custom configuration file.
--debug Debug logging.
--debug-grpc Debug gRPC logging. Very verbose, used for debugging connection problems.
--no-user-output Disable printing user intended output to stderr.
--retry int
Enable gRPC retries. By default, retries are enabled with maximum 5 attempts.
Pass 0 to disable retries. Pass any negative value for infinite retries.
Even infinite retries are capped with 2 minutes timeout.
--cloud-id string
Set the ID of the cloud to use.
--folder-id string
Set the ID of the folder to use.
--folder-name string
Set the name of the folder to use (will be resolved to id).
--endpoint string
Set the Cloud API endpoint (host:port).
--token string
Set the OAuth token to use.
--impersonate-service-account-id string
Set the ID of the service account to impersonate.
--no-browser Disable opening browser for authentication.
--format string
Set the output format: text (default), yaml, json, json-rest.
--jq string
Query to select values from the response using jq syntax
-h,--help Display help for the command.
Previous
add-node-labels
Next
delete