Skip to content
Xinlian Liu edited this page Aug 9, 2019 · 1 revision

Connect to the Enclave

  1. Connect to LBL VPN

  2. either https://vavdi.ornl.gov in browser and use the HTML client, or use the VMWare Horizon Client to connect to the same address.

  3. Type in your passcode (six-digit personal PIN + RSA token) to log in the enclave

  4. Connect to the Windows VM (mvp011-01) or either of the two Linux VMs (mvp011-02 and mvp011-03)

  5. Login a VM with your KDI user name ([email protected]) and password, which will be set after you log in with the temporary password for the first time.

Connect to VASQL

  • From Windows using Microsoft SQL Server Management Studio (SSMS):

    1. Launch Windows PowerShell
    2. Type SSMS.exe
  • From Linux through sqlcmd:

    Example:

    $ /opt/mssql-tools/bin/sqlcmd -D -SMVPSQL -Q "SELECT TOP 10 * FROM CDWWork.Patient.Patient"
  • From Linux through Azure Data Studio:

    Example:

    $ azuredatastudio

    Use VASQL as database name and choose Windows Authentication.

Data Processing

Read RDS files prepared by Ben McMahon:

>> import pyreadr
>> results = pyreadr.read_r('demo42.RDS')

The HPC Environment

Computer resources

VM

  • Linux VM: Xeon Gold 6142, 8 threads, 64GB RAM, 100GB storage
  • Windows limts maximum concurrent users to 2

The Computer cluster:

  • 30 CPU nodes, 36 CPU cores, 128 GB RAM, 2TB HDD
  • 15 GPU nodes, dual K80 GPUs, 36 CPU cores, 128 GB RAM, 2 TB HDD
  • 3.2 PB Lustre file system

  • 5 PB NFS and tape file system

The Relational Database

  • 2 nodes, 40 cores, 3TB RAM, 11TB SSD storage, 167 TB HHD storage

System Repositories:

  • CentOS 7 Base, CentOS 7 Updates, CentOS 7 Extras, EPEL

User Repositories:

  • Anaconda, install by /usr/local/bin/condaenv
  • CRAN
  • PyPI

VM Project Shared folder 1.5 TB:

  • Windows: file:\\vafs01\data\VA_MVP011_CDW

  • Linux: /mnt/vafs01/data/

Murphy:

  • project shared folder: /mnt/vafs01/data, available on login node only
  • HPC shared folder: /lustre/valustre/VA_MVP011_CDW
  • Local scratch: /localscratch

Connect to HPC

$ ssh va-murphy-login.kdi.local

Install Jupyter Notebook on Linux VM

$ python -m pip install jupyter

Install Anaconda on Linux VM

$ condaenv

Setting anaconda path will conflict with Linux VM remote desktop. Do not let the installer change the .bashrc file.

Using Jupyter Notebook on Murphy

From the Murphy Login node:

$ module load apps/python/anaconda/3/4.3.1
$ jupyter notebook --no-browser --ip=0.0.0.0

From VM:

$ ssh -N -f -L locahost:8889:localhost:8888 va-murphy-login.kdi.local

Type in a browser search bar: locahost:8889

Request new software

EMail [email protected], with

Subject: VA Tenant Software Change Request

Research Project Title:

Project Technical Contact:

Requestor Contact Information

Software Name:

Software Version(s), or Git revision/ SVN revision/Release updates:

Release Date:

License Agreement:

Software Homepage URL:

Installation Target OS:

Installation target(s) in KDI: specific VM, or cluster names

Prerequisites/dependencies:

Purpose/Justification:

Documentation

GitLab

https://code-va.kdi.local

Use Ad Login and your userid liux.mvp011w without the kdi.local part