These are commonly reported issues when you work with Kubernetes.
Kubernetes requires Docker to be running on the machine. Our Kubernetes blueprint attempts to install Docker on your VMs. However, installation may time out. This usually indicates that it failed to install Docker.
2018-05-01 09:16:46.272 CFY <kubernetes> [kubernetes_master_y032hn.create] Sending task 'script_runner.tasks.run' 2018-05-01 09:16:46.272 CFY <kubernetes> [kubernetes_master_y032hn.create] Task started 'script_runner.tasks.run' 2018-05-01 09:16:46.410 LOG <kubernetes> [kubernetes_master_y032hn.create] INFO: Downloaded scripts/kubernetes_master/create.py to /tmp/M2Z4A/tmpH2xJaV-create.py 2018-05-01 09:16:47.276 CFY <kubernetes> [kubernetes_master_y032hn.create] Task rescheduled 'script_runner.tasks.run' -> Docker is not present on the system.
Troubleshooting this issue depends on which IaaS you use:
* The AWS blueprint uses Cloud Init to install Docker. Make sure that your AMI supports Cloud Init. If so, check the Cloud Init log `/var/log/cloud-init.log`. * The Openstack blueprint uses Cloud Init to install Docker. Make sure that your Openstack image supports Cloud Init. If so, check the Cloud Init log `/var/log/cloud-init.log`. * The Azure blueprint uses Custom Scripts. Check the [troubleshooting guide](https://docs.microsoft.com/en-us/azure/virtual-machines/linux/extensions-customscript#troubleshooting) for Custom Scripts. * The GCP blueprint uses Startup Scripts. See the documentation on [Startup Scripts](https://cloud.google.com/compute/docs/startupscript). * The VSphere blueprint uses the Cloudify script plugin. The deployment execution logs should be helpful.
Kubernetes uses client certificate authentication. These certificates are base64-encoded.
File "/usr/lib64/python2.7/base64.py", line 321, in decodestring return binascii.a2b_base64(s) Error: Incorrect padding
This error is shown when the encoded value is corrupted. You can try to decode and re-encode the certificate.