deployDSVMCluster: Deploy a cluster of Linux Data Science Virtual Machines on...

In Azure/AzureDSVM: Create and Manage Azure Data Science Virtual Machine(s) for Convenient Execution of Scalable and Elastic Analytics.

Description

Creates a cluster of Data Science Virtual Machines and enables the DSVMs to communicate acros the cluster via public key based credentials for high performance computation. All DSVMs in the cluster are based on Linux OS and use public key cryptogrphy for log in.

Usage

deployDSVMCluster(context, resource.group, location, hostname, username,
  authen = "Key", pubkey = "", password = "", os = "Ubuntu",
  size = "Standard_D1_v2", dns.label = hostname, count)

Arguments

context	Authentication context of AzureSMR encapsulating the TID, CID, and key obtained from Azure Actrive Directory.
resource.group	The Azure resource group where the DSVM is created.
location	Location of the data centre to host the DSVM.
hostname	Name of the DSVM. Lowercase characters or numbers only. Special characters are not permitted.
username	User name of the DSVM. It should be different from 'name'.
authen	Either "Key" for public-key based authentication (with Linux) or "Password" for a password based authentication (Linux or Windows). Default is to use public key authentication for Linux and password based authentication for Windows.
pubkey	Public key for the DSVM. Only applicable for public-key based authentication of Linux based DSVM.
password	Pass word for the DSVM.
os	Operating system of DSVM. Permitted values are "Ubuntu", "CentOS", "Windows", "DeepLearning", and "MRS". The default is to deploy a Ubuntu Linux Data Science Virtual Machine. NOTE previously Windows based DSVM for deep learning toolkit is using a separated package but now it is merged into one.
size	Size of the DSVM. The default is "Standard_D1_v2". All available sizes can be obtained by function 'getVMSizes'.
dns.label	DNS label for the VM address. The URL for accessing the deployed DSVM will be "<dns_label>.<location>.cloudapp.azure.com
count	If provided this is the number of DSVM instances to be created. If not provided the number of DSVMs created will be either the number of names provided or the number of username provided.

Details

Note clustering of DSVMs for high performance computing will be deprecated. Users are encouraged to use azureDoParallel package for the same purpose with Azure Batch Services.

The current function supports for deployment of identical DSVMs (i.e., size, OS, etc.) or heterogeneous DSVMs with differerent specifications.

Examples

## Not run: 
# The following deploys a cluster of 5 Linux DSVMs.

deployDSVMCluster(context, resource.group="<resource_group>", 
location="<location>", hostname="<machine_name>", username="<user_name>", 
os="Windows", pubkey="<a_valid_public_key_string_in_SSH_format>", count=5)

# The following deploys a collection of 3 Linux DSVMs with different 
names and public keys.

deployDSVMCluster(context, resource.group="<resource_group>", 
location="<location>", hostname=c("<machine_name_1>", "<machine_name_2>", 
"<machine_name_3>", username=c("<user_name_1>", "<user_name_2>", 
"<user_name_3>"), os="Windows", 
pubkey="<a_valid_public_key_string_in_SSH_format>"
## End(Not run)

Azure/AzureDSVM documentation built on May 20, 2019, 2:03 p.m.