If you see:

$ pcluster create mycluster
Creating stack named: parallelcluster-mycluster
Status: parallelcluster-mycluster - ROLLBACK_IN_PROGRESS                        
Cluster creation failed.  Failed events:
  - AWS::EC2::Instance MasterServer Received FAILURE signal with UniqueId i-07af1cb218dd6a081

There was a problem creating the cluster, to diagnose, re-run the create with the --norollback flag. Then ssh into the cluster:

$ pcluster create mycluster --norollback
...
$ pcluster ssh mycluster

From there, there's three log files that will tell you what error occurred:

• /var/log/cfn-init.log start here, likely you'll see an error like Command chef failed, look above that line for the specific error
• /var/log/cloud-init.log if you don't see anything telling in cfn-init.log this a good resource
• /var/log/cloud-init-output.log you can view the stdout of the cfn-init command. Usually not needed.

Affecting pcluster versions <= 2.1.1

If in the /var/log/cfn-init.log file you see this,

STDERR: Traceback (most recent call last):
  File "/usr/local/sbin/attachVolume.py", line 90, in <module>
    main()
  File "/usr/local/sbin/attachVolume.py", line 68, in main
    response = ec2.attach_volume(VolumeId=volumeId, InstanceId=instanceId, Device=dev)
  File "/usr/lib/python2.7/site-packages/botocore/client.py", line 357, in _api_call
    return self._make_api_call(operation_name, kwargs)
  File "/usr/lib/python2.7/site-packages/botocore/client.py", line 661, in _make_api_call
    raise error_class(parsed_response, operation_name)
botocore.exceptions.ClientError: An error occurred (InvalidParameterValue) when calling the AttachVolume operation: Invalid value '/dev/sdb' for unixDevice. Attachment point /dev/sdb is already in use

This is a known bug in pcluster-2.1.1 that effects NVME based instances type, such as c5, m5, z1d. See #823 for more info.

Upgrade to a version > 2.1.1 to fix.

If you see:

$ pcluster create mycluster
Beginning cluster creation for cluster: mycluster
Creating stack named: parallelcluster-mycluster
Status: parallelcluster-mycluster - CREATE_FAILED                               
Cluster creation failed.  Failed events:
  - AWS::CloudFormation::Stack parallelcluster-mycluster The following resource(s) failed to create: [ComputeFleet]. 
  - AWS::AutoScaling::AutoScalingGroup ComputeFleet You must use a valid fully-formed launch template. The requested configuration is currently not supported. Please check the documentation for supported configurations. (Service: AmazonAutoScaling; Status Code: 400; Error Code: ValidationError; Request ID: 1737665d-2b08-11e9-9141-df273bd3069e)

This issue means the requested instance type isn't available in the region you requested. This can either be due to insufficient capacity in that region / availability zone, or it's unable to get the instances you need in the placement group you requested (see 3).

1. Change placement group by changing the compute_subnet_id or, if you only use one subnet, master_subnet_id.
2. Try a different region by changing the aws_region_name = in your ~/.parallelcluster/config
3. If you're using a placement group, change placement = compute in your config. It defaults to this as of 2.2.0. By requesting a placement group with different head node and compute instances types, it becomes less likely that those instances are co-located, and hence the instance launch will fail to find the instances it needs to launch the cluster. Setting placement = compute changes this, by only putting the compute instances in the placement group.