cluster creation with existing FSx throws traceback

[wleepang]

Environment:

• AWS ParallelCluster :aws-parallelcluster-2.2.1
• OS: alinux
• Scheduler: SGE
• Master instance type: t2.micro
• Compute instance type: t2.micro

Bug description and how to reproduce:
Creating a cluster with a pre-existing FSx file system fails with a traceback:

[ec2-user@ip-172-31-41-28 ~]$ pcluster create cromwell
Beginning cluster creation for cluster: cromwell
Traceback (most recent call last):
  File "/usr/bin/pcluster", line 11, in <module>
    load_entry_point('aws-parallelcluster==2.2.1', 'console_scripts', 'pcluster')()
  File "/usr/lib/python2.7/site-packages/pcluster/cli.py", line 354, in main
    args.func(args)
  File "/usr/lib/python2.7/site-packages/pcluster/cli.py", line 27, in create
    pcluster.create(args)
  File "/usr/lib/python2.7/site-packages/pcluster/pcluster.py", line 76, in create
    config = cfnconfig.ParallelClusterConfig(args)
  File "/usr/lib/python2.7/site-packages/pcluster/cfnconfig.py", line 99, in __init__
    self.__init_fsx_parameters()
  File "/usr/lib/python2.7/site-packages/pcluster/cfnconfig.py", line 827, in __init_fsx_parameters
    self.__validate_resource("fsx_fs_id", (value, self.__master_subnet))
  File "/usr/lib/python2.7/site-packages/pcluster/cfnconfig.py", line 281, in __validate_resource
    self.__resource_validator.validate(resource_type, resource_value)
  File "/usr/lib/python2.7/site-packages/pcluster/config_sanity.py", line 517, in validate
    self.__validate_fsx_parameters(resource_type, resource_value)
  File "/usr/lib/python2.7/site-packages/pcluster/config_sanity.py", line 195, in __validate_fsx_parameters
    self.__check_fsx_fs_id(ec2, fsx, resource_value)
  File "/usr/lib/python2.7/site-packages/pcluster/config_sanity.py", line 173, in __check_fsx_fs_id
    "inbound and outbound TCP traffic from 0.0.0.0/0 through port 988." % resource_value[0]
TypeError: __fail() takes exactly 2 arguments (1 given)

Additional context:

Config file is:

[aws]

[cluster cromwell]
vpc_settings = public
key_name = ********
initial_queue_size = 2
maintain_initial_size = true
fsx_settings = fs

[fsx fs]
shared_dir = /cromwell_root
fsx_fs_id = fs-0f0ddbaf5d3781422

[vpc public]
master_subnet_id = subnet-c264e1a5
vpc_id = vpc-461dcc3c

[global]
update_check = true
sanity_check = true
cluster_template = cromwell

[aliases]
ssh = ssh {CFN_USER}@{MASTER_IP} {ARGS}

The FSx Lustre file system is in us-east-1 where the cluster is created.

[sean-smith]

@wleepang This is because the default base_os = alinux which isn't currently supported. Fix it by switching to centos:

[cluster cromwell]
...
base_os = centos7

I know this is confusing, so I wrote validation to check before cluster create, in #904 but it hasn't been released yet.

Also stayed tuned for amazon linux support in the very near future ;-).

[wleepang]

@sean-smith - Thanks! I did eventually figure this out by reading a blog by Jiawei Zhuang. Looking forward to the update!