diff options
Diffstat (limited to 'TOSCA/kubernetes-cluster-TOSCA/scripts/kubernetes_node')
-rw-r--r-- | TOSCA/kubernetes-cluster-TOSCA/scripts/kubernetes_node/configure.py | 88 |
1 files changed, 88 insertions, 0 deletions
diff --git a/TOSCA/kubernetes-cluster-TOSCA/scripts/kubernetes_node/configure.py b/TOSCA/kubernetes-cluster-TOSCA/scripts/kubernetes_node/configure.py new file mode 100644 index 0000000000..69faaa80d1 --- /dev/null +++ b/TOSCA/kubernetes-cluster-TOSCA/scripts/kubernetes_node/configure.py @@ -0,0 +1,88 @@ +#!/usr/bin/env python + +# ============LICENSE_START========================================== +# =================================================================== +# Copyright © 2017 AT&T +# +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. +#============LICENSE_END============================================ + +# Afther K8s master up and running. This script will be triggered in each worker nodes. It will join the nodes, and mount the NFS directory. + +import subprocess +from cloudify import ctx +from cloudify.exceptions import NonRecoverableError + +START_COMMAND = 'sudo kubeadm join --token {0} {1}:{2}' + + +def execute_command(_command): + + ctx.logger.debug('_command {0}.'.format(_command)) + + subprocess_args = { + 'args': _command.split(), + 'stdout': subprocess.PIPE, + 'stderr': subprocess.PIPE + } + + ctx.logger.debug('subprocess_args {0}.'.format(subprocess_args)) + + process = subprocess.Popen(**subprocess_args) + output, error = process.communicate() + + ctx.logger.debug('command: {0} '.format(_command)) + ctx.logger.debug('output: {0} '.format(output)) + ctx.logger.debug('error: {0} '.format(error)) + ctx.logger.debug('process.returncode: {0} '.format(process.returncode)) + + if process.returncode: + ctx.logger.error('Running `{0}` returns error.'.format(_command)) + return False + + return output + + +if __name__ == '__main__': + + hostname = execute_command('hostname') + ctx.instance.runtime_properties['hostname'] = hostname.rstrip('\n') + + # Get the master cluster info. + masters = \ + [x for x in ctx.instance.relationships if + x.target.instance.runtime_properties.get( + 'KUBERNETES_MASTER', False)] + if len(masters) != 1: + raise NonRecoverableError( + 'Currently, a Kubernetes node must have a ' + 'dependency on one Kubernetes master.') + master = masters[0] + bootstrap_token = \ + master.target.instance.runtime_properties['bootstrap_token'] + master_ip = \ + master.target.instance.runtime_properties['master_ip'] + master_port = \ + master.target.instance.runtime_properties['master_port'] + + # Join the cluster. + cniCommand1=subprocess.Popen(["sudo", "sysctl", 'net.bridge.bridge-nf-call-iptables=1'], stdout=subprocess.PIPE) + join_command = \ + 'sudo kubeadm join --token {0} {1}:{2}'.format( + bootstrap_token, master_ip, master_port) + execute_command(join_command) + + #mount + mount_command=\ + 'sudo mount -t nfs -o proto=tcp,port=2049 {0}:/dockerdata-nfs /dockerdata-nfs'.format(master_ip) + execute_command(mount_command)
\ No newline at end of file |