subctl

subctl is a command line utility designed to simplify the deployment and maintenance of Submariner across your clusters.

subctl helps to automate the deployment of the Submariner Operator, thereby reducing the possibility of mistakes during the process.

subctl connects to specified cluster(s) and performs the requested command.

Synopsis

subctl [command] [--flags] ...

Installation

Download the subctl binary and make it available on your PATH. 

curl -Ls https://get.submariner.io | bash
export PATH=$PATH:~/.local/bin
echo export PATH=\$PATH:~/.local/bin >> ~/.profile

Installing specific versions

By default, https://get.submariner.io will provide the latest release for subctl, and hence Submariner. Specific versions can be requested by using the VERSION environment variable.

Avalailable options are:

  • latest: the latest stable release (default)
  • devel: the master branch code.
  • rc: the latest release candidate.
  • x.x.x (like 0.6.1, 0.5.0, etc)

For example

curl https://get.submariner.io | VERSION=devel bash

Commands

deploy-broker

subctl deploy-broker [flags]

The deploy-broker command configures the cluster specified by the --kubeconfig flag (or KUBECONFIG env var) and the --kubecontext flag as the Broker. It installs the necessary CRDs and the submariner-k8s-broker namespace.

In addition, it generates a broker-info.subm file which can be used with the join command to connect clusters to the Broker. This file contains the following details:

  • Encryption PSK key
  • Broker access details for subsequent subctl runs
  • Service Discovery settings

deploy-broker flags

Flag Description
--kubeconfig <string> Absolute path(s) to the kubeconfig file(s) (default $HOME/.kube/config)
--kubecontext <string> kubeconfig context to use
--service-discovery Enable Multi Cluster Service Discovery (default true)
--globalnet Enable support for overlapping Cluster/Service CIDRs in connecting clusters (default disabled)
--globalnet-cidr-range <string> Global CIDR supernet range for allocating GlobalCIDRs to each cluster (default “169.254.0.0/16”)
--ipsec-psk-from <string> Import IPsec PSK from existing Submariner broker file, like broker-info.subm (default broker-info.subm)

export

export service

subctl export service [flags] <name> creates a ServiceExport resource for the given Service name. This makes the corresponding Service discoverable from other clusters in the Submariner deployment.

export service flags

Flag Description
--kubeconfig <string> Absolute path(s) to the kubeconfig file(s) (default $HOME/.kube/config)
--kubecontext <string> Kubeconfig context to use
--namespace <string> Namespace in which the Service to be exported belongs

If no namespace flag is specified, it uses the default namespace from the current context, if present, otherwise it uses default.

join

subctl join broker-info.subm [flags]

The join command deploys the Submariner Operator in a cluster using the settings provided in the broker-info.subm file. The service account credentials needed for the new cluster to access the Broker cluster will be created and provided to the Submariner Operator deployment.

join flags (general)

Flag Description
--cable-driver <string> Cable driver implementation. Available options are libreswan (default), strongswan, and wireguard
--clusterid <string> Cluster ID used to identify the tunnels. Every cluster needs to have a unique cluster ID. If not provided, one will be generated by default based on the cluster name in the kubeconfig file
--clustercidr <string> Specifies the cluster’s CIDR used to generate Pod IP addresses. If not specified, subctl will try to discover it and if unable to do so, it will prompt the user
--no-label Skip gateway labeling. This disables the prompt for a Worker node to use as gateway
--enable-pod-debugging Enable Submariner pod debugging (verbose logging in the deployed pods)

join flags (Globalnet)

Flag Description
--globalnet-cluster-size <value> Cluster size for GlobalCIDR allocated to this cluster (amount of global IPs)
--globalnet-cidr <string> GlobalCIDR to be allocated to the cluster, this setting is exclusive with --globalnet-cluster-size and configures a specific Globalnet CIDR for this cluster

join flags (IPsec)

Flag Description
--disable-nat Disable NAT for IPsec
--ikeport <value> IPsec IKE port (default 500)
--ipsec-debug Enable IPsec debugging (verbose logging)
--nattport <value> IPsec NAT-T port (default 4500)

join flags (images and repositories)

Flag Description
--repository <string> The repository from where the various Submariner images will be sourced (default “quay.io/submariner”)
--version <string> Image version
--image-override <string>=<string> Component image override. This flag can be used more than once (example: –image-override=submariner=quay.io/myUser/submariner:latest)

show

show networks

subctl show networks [flags]

Inspects the cluster and reports information about the detected network plugin and detected Cluster and Service CIDRs.

show versions

subctl show versions [flags]

Shows the version and image repository of each Submariner component in the cluster.

show gateways

subctl show gateways [flags]

Shows summary information about the Submariner gateways in the cluster.

show connections

subctl show connections [flags]

Shows information about the Submariner endpoint connections with other clusters.

show endpoints

subctl show endpoints [flags]

Shows information about the Submariner endpoints in the cluster.

show all

subctl show all [flags]

Shows the aggregated information from all the other show commands.

show flags

Flag Description
--kubeconfig <string> Absolute path(s) to the kubeconfig file(s) (default $HOME/.kube/config)
--kubecontext <string> Kubeconfig context to use

verify

subctl verify <kubeConfig1> <kubeConfig2> [flags]

The verify command verifies a Submariner deployment between two clusters is functioning properly. The kubeConfig1 file will be ClusterA in the reports, while kubeConfig2 will be ClusterB in the reports. The --verbose flag is recommended to see what’s happening during the tests.

There are several suites of verifications that can be performed. By default all verifications are performed. Some verifications are deemed disruptive in that they change some state of the clusters as a side effect. If running the command interactively, you will be prompted for confirmation to perform disruptive verifications unless the --enable-disruptive flag is also specified. If running non-interactively (that is with no stdin), --enable-disruptive must be specified otherwise disruptive verifications are skipped.

The connectivity suite verifies dataplane connectivity across the clusters for the following cases:

  • Pods (on Gateway nodes) to Services
  • Pods (on non-Gateway nodes) to Services
  • Pods (on Gateway nodes) to Pods
  • Pods (on non-Gateway nodes) to Pods

The service-discovery suite verifies DNS discovery of <service>.<namespace>.svc.clusterset.local entries across the clusters.

The gateway-failover suite verifies the continuity of cross-cluster dataplane connectivity after a gateway failure in a cluster occurs. This suite requires a single gateway configured on ClusterA and other available Worker nodes capable of serving as gateways. Please note that this verification is disruptive.

verify flags

Flag Description
--connection-attempts <value> The maximum number of connection attempts (default 2)
--connection-timeout <value> The timeout in seconds per connection attempt (default 60)
--operation-timeout <value> Operation timeout for Kubernetes API calls (default 240)
--report-dir <string> XML report directory (default “.")
--verbose Produce verbose logs during connectivity verification
--only Comma separated list of specific verifications to perform
--enable-disruptive Enable verifications which are potentially disruptive to your deployment

benchmark

benchmark throughput

subctl benchmark throughput <kubeconfig1> [<kubeconfig2>] [flags]

The benchmark throughput command runs a throughput benchmark test between two specified clusters or within a single cluster. It deploys a Pod to run the iperf tool and logs the output to the console. When running benchmark throughput, two types of tests will be executed:

  • Pod to Pod - where both Pods are scheduled on Gateway nodes
  • Pod to Pod - where both Pods are scheduled on non-Gateway nodes

benchmark latency

subctl benchmark latency <kubeconfig1> [<kubeconfig2>] [flags]

The benchmark latency command runs a latency benchmark test between two specified clusters or within a single cluster. It deploys a Pod to run the netperf tool and logs the output to the console. When running benchmark latency, two types of tests will be executed:

  • Pod to Pod - where both Pods are scheduled on Gateway nodes
  • Pod to Pod - where both Pods are scheduled on non-Gateway nodes

benchmark flags

Flag Description
--intra-cluster Performs the benchmark test within a single cluster between Pods from a Non-Gateway node to a Gateway node
--verbose Produce verbose logs during benchmark tests

version

subctl version

Prints the version details for the subctl binary.