Building and running SkyLab

List of spack, software, and AMIs

Versions used:

Note. It is necessary to use c6i.4xlarge or larger instances of this family (recommended: c6i.8xlarge when running the skylab-atm-land-small experiment).

For more information about using Amazon Web Services please see JEDI on AWS.

Developer section

Note. To follow this section, one needs read access to the JCSDA-internal GitHub org.

1- Load modules

First, you need to load all the modules needed to build jedi-bundle and solo/r2d2/ewok/simobs/skylab. Note that loading modules only sets up the environment for you. You still need to build jedi-bundle, run ctests, install solo/r2d2/ewok/simobs and download skylab.

Please note that currently we only support Orion, Hercules, Derecho, Discover, S4, and AWS platforms. If you are working on a system not specified below please follow the instructions on JEDI Portability.

Users are responsible for setting up their GitHub and AWS credentials on the platform they are using. You will need to create or edit your ~/.aws/config and ~/.aws/credentials to make sure they contain:

[default]
region=us-east-1

# NOAA AWS account configuration for the ``jcsda-noaa-aws-us-east-1`` R2D2 Data Hub
[jcsda-noaa-aws-us-east-1]
region=us-east-1

# USAF AWS account configuration for the ``jcsda-usaf-aws-us-east-2`` R2D2 Data Hub
[jcsda-usaf-aws-us-east-2]
region=us-east-2

# NOAA AWS account credentials if default in config is us-east-1
[default]
aws_access_key_id=***
aws_secret_access_key=***

# NOAA AWS account credentials for the ``jcsda-noaa-aws-us-east-1`` R2D2 Data Hub
[jcsda-noaa-aws-us-east-1]
aws_access_key_id=***
aws_secret_access_key=***

# USAF AWS account credentials for the ``jcsda-usaf-aws-us-east-2`` R2D2 Data Hub
[jcsda-usaf-aws-us-east-2]
aws_access_key_id=***
aws_secret_access_key=***

Tip

Make sure to protect your AWS config and credentials via:

chmod 400 ~/.aws/config
chmod 400 ~/.aws/credentials

The commands for loading the modules to compile and run SkyLab are provided in separate sections for HPC platforms and AWS instances (AMIs). Users need to execute these commands before proceeding with the build of jedi-bundle below.

Warning

If you are using spack-stack 1.4.0 or spack-stack 1.4.1 you need to unload the CRTM v2.4.1-jedi module after loading the Spack-Stack modules.

module unload crtm

Make sure you are building CRTMV3 within the jedi-bundle using the ecbuild_bundle command.

2- Build jedi-bundle

Once the stack is installed and the corresponding modules loaded, the next step is to get and build the JEDI executables.

The first step is to create your work directory. In this directory you will clone the JEDI code and all the files needed to build, test, and run JEDI and SkyLab. We call this directory JEDI_ROOT throughout this document.

The next step is to clone the code bundle to a local directory:

mkdir $JEDI_ROOT
cd $JEDI_ROOT
git clone https://github.com/jcsda/jedi-bundle

The example here is for jedi-bundle, the instructions apply to other bundles as well.

From this point, we will use two environment variables:

  • $JEDI_SRC which should point to the base of the bundle to be built (i.e. the directory that was cloned just above, where the main CMakeLists.txt is located or $JEDI_ROOT/jedi-bundle). $JEDI_SRC=$JEDI_ROOT/jedi-bundle

  • $JEDI_BUILD which should point to the build directory or $JEDI_ROOT/build. Create the directory if it does not exist. $JEDI_BUILD=$JEDI_ROOT/build

Note:

It is recommended these two directories are not one inside the other.

  • Orion: it’s recommended to use $JEDI_ROOT=/work2/noaa/jcsda/${USER}/jedi.

  • Discover: it’s recommended to use $JEDI_ROOT=/discover/nobackup/${USER}/jedi.

  • On AWS Parallel Cluster, use $JEDI_ROOT=/mnt/experiments-efs/USER.NAME/jedi.

  • On the preconfigured AWS AMIs, use $JEDI_ROOT=$HOME/jedi.

Building JEDI then can be achieved with the following commands:

mkdir $JEDI_BUILD
cd $JEDI_BUILD
ecbuild $JEDI_SRC
make -j8

Feel free to have a coffee while it builds. Once JEDI is built, you should check the build was successful by running the tests (still from $JEDI_BUILD):

ctest

If you are on an HPC you may need to provide additional flags to the ecbuild command, or login to a compute node, or submit a batch script for running the ctests. Please refer the Skylab HPC users guide for more details.

(You might have another coffee.) You have successfully built JEDI!

Warning

Even if you are a master builder and don’t need to check your build, if you intend to run experiments with ewok, you still need to run a few of the tests that download data (this is temporary). You can run these tests with:

ctest -R get_

If you are running on your own machine you will also need to clone the static-data repo for some skylab experiments.

cd $JEDI_SRC
git clone https://github.com/jcsda-internal/static-data

Note

Run ctest --help for more information on the test options. Also, you may find yourself in a situation in which you only want to run a single test such as test_soca_lefkf and see its verbose output, but excecuting ctest -VV -R test_soca_lefkf will also run and write the output for the test_soca_letkf_split_observer and test_soca_letkf_split_solver tests. To run only the first test enter a $ at the end of the test name like this:

ctest -VV -R test_soca_lefkf$

3- Clone and install solo/r2d2/ewok/simobs, clone skylab only

We recommend that you use a python3 virtual environment (venv) for building solo/r2d2/ewok/simobs. As indicated above in the note about the $JEDI_SRC and $JEDI_BUILD environment variables, clone these repos inside the clone of the jedi-bundle repo.

cd $JEDI_SRC
git clone https://github.com/jcsda-internal/solo
git clone https://github.com/jcsda-internal/r2d2
git clone https://github.com/jcsda-internal/ewok
git clone https://github.com/jcsda-internal/simobs
git clone https://github.com/jcsda-internal/skylab

Or for the latest release of Skylab v7, clone the corresponding workflow repository branches:

cd $JEDI_SRC
git clone --branch 1.2.0 https://github.com/jcsda-internal/solo
git clone --branch 2.3.0 https://github.com/jcsda-internal/r2d2
git clone --branch 0.7.0 https://github.com/jcsda-internal/ewok
git clone --branch 1.5.0 https://github.com/jcsda-internal/simobs
git clone --branch 7.0.0 https://github.com/jcsda-internal/skylab

Continue with setting up a virtual environment.

cd $JEDI_ROOT
python3 -m venv --system-site-packages venv
source venv/bin/activate

You can then proceed with

cd $JEDI_SRC/solo
python3 -m pip install -e .
cd $JEDI_SRC/r2d2
python3 -m pip install -e .
cd $JEDI_SRC/ewok
python3 -m pip install -e .
cd $JEDI_SRC/simobs
python3 -m pip install -e .

Note: You need to run source venv/bin/activate every time you start a new session on your machine.

4- Setup SkyLab

Create and source $JEDI_ROOT/activate.sh

We recommend creating this bash script and sourcing it before running the experiment. This bash script sets environment variables such as JEDI_BUILD, JEDI_SRC, EWOK_WORKDIR and EWOK_FLOWDIR required by ewok. If these variables are not defined they will be set from JEDI_ROOT.

Users may set JEDI_SRC, JEDI_BUILD, EWOK_WORKDIR and EWOK_FLOWDIR to point to relevant directories on their systems or use the default template in the sample script below. Note that JEDI_SRC, JEDI_BUILD and EWOK_WORKDIR are experiment specific, i.e. you can run several experiments at the same time, each having their own definition for these variables.

The user further has to set two environment variables R2D2_HOST and R2D2_COMPILER in the script. R2D2_HOST and R2D2_COMPILER are required by r2d2 and ewok. They are used to initialize the location EWOK_STATIC_DATA of the static data used by skylab and bind r2d2 to your current environment. EWOK_STATIC_DATA is staged on the preconfigured platforms.

In the section that exports your R2D2_HOST and R2D2_COMPILER, Be sure to remove all lines that are NOT relevant to your platform.

On generic platforms, the script sets EWOK_STATIC_DATA to ${JEDI_SRC}/static-data/static.

Please don’t forget to source this script after creating it: source $JEDI_ROOT/activate.sh

Please see Skylab HPC users guide for more information on specifics for editing this activate.sh script and other general instructions and notes for running skylab on supported HPC systems.

#!/bin/bash

# Set JEDI_ROOT

if [ -z $JEDI_ROOT ]; then
  export JEDI_ROOT=**Set this based on your set up if JEDI_SRC, JEDI_BUILD, EWOK_WORKDIR or EWOK_FLOWDIR are not defined.**
fi

if [ -z $JEDI_SRC ]; then
  export JEDI_SRC=${JEDI_ROOT}/jedi-bundle
fi

# Set the host for R2D2/EWOK

# On Derecho
export R2D2_HOST=derecho
# On Discover
export R2D2_HOST=discover
# On Hercules
export R2D2_HOST=hercules
# On Orion
export R2D2_HOST=orion
# On S4
export R2D2_HOST=s4
# On AWS Parallel Cluster
export R2D2_HOST=aws-pcluster
# On NOAA's ParallelWorks on AWS
export R2D2_HOST=pw-aws
# On NOAA's ParallelWorks on Azure
export R2D2_HOST=pw-azure
# On NOAA's ParallelWorks on GCloud
export R2D2_HOST=pw-gcloud
# On your local machine / AWS single node
export R2D2_HOST=localhost

# Set the compiler for R2D2/EWOK

# For gnu
export R2D2_COMPILER=gnu
# For intel
export R2D2_COMPILER=intel
# For clang/llvm
export R2D2_COMPILER=clang

# Most users won't need to change the following settings

# Source source this file for ewok ecFlow workflows
source $JEDI_ROOT/venv/bin/activate

if [ -z $JEDI_BUILD ]; then
  export JEDI_BUILD=${JEDI_ROOT}/build
fi

if [ -z $EWOK_WORKDIR ]; then
  export EWOK_WORKDIR=${JEDI_ROOT}/workdir
fi

if [ -z $EWOK_FLOWDIR ]; then
  export EWOK_FLOWDIR=${JEDI_ROOT}/ecflow
fi

# Add ioda python bindings to PYTHONPATH
PYTHON_VERSION=`python3 -c 'import sys; version=sys.version_info[:2]; print("{0}.{1}".format(*version))'`
export PYTHONPATH="${JEDI_BUILD}/lib/python${PYTHON_VERSION}:${PYTHONPATH}"

# necessary user directories for ewok and ecFlow files
mkdir -p $EWOK_WORKDIR $EWOK_FLOWDIR

# ecFlow vars - EDIT: (set to constant) if running locally or on Discover
myid=$(id -u ${USER})
if [[ $myid -gt 64000 ]]; then
  myid=$(awk -v min=3000 -v max=31000 -v seed=$RANDOM 'BEGIN{srand(seed); print int(min + rand() * (max - min + 1))}')
fi
export ECF_PORT=$((myid + 1500))

# The ecflow hostname (e.g. a specific login node) is different from the R2D2/EWOK general host (i.e. system) name
host=$(hostname | cut -f1 -d'.')
export ECF_HOST=$host

case $R2D2_HOST in
  derecho)
    export EWOK_STATIC_DATA=/glade/p/mmm/jedipara/static
    ;;
  discover)
    export EWOK_STATIC_DATA=/discover/nobackup/projects/jcsda/s2127/static
    ;;
  orion* | hercules)
    export EWOK_STATIC_DATA=/work/noaa/da/role-da/static
    ;;
  s4)
    export EWOK_STATIC_DATA=/data/prod/jedi/static
    ;;
  aws-pcluster)
    export EWOK_STATIC_DATA=${JEDI_ROOT}/static
    ;;
  pw-aws | pw-azure | pw-gcloud)
    echo "EWOK_STATIC_DATA chas not been assigned to '$R2D2_HOST'"
    ;;
  localhost)
    export EWOK_STATIC_DATA=${JEDI_SRC}/static-data/static
    ;;
  *)
    echo "Unknown compute host name '$R2D2_HOST'"
    exit 1
    ;;
esac

If you are running locally you my want to pick a constant value for ECF_PORT. As written, the code above will generate a new, random value for your ECF_PORT everytime this script is sourced. Changing your ECF_PORT will require you to reconnect the ecflow server after everytime this script is sourced, so keeping it constant will keep your ecflow server connected.

Note: On AWS pcluster users will need to update the python version referenced in the above source $JEDI_ROOT/activate.sh script. The following lines under # ecflow and pyioda Python bindings should be:

# ecflow and pyioda Python bindings
PYTHON_VERSION=`python3 -c 'import sys; version=sys.version_info[:2]; print("{0}.{1}".format(*version))'`
export PYTHONPATH="${JEDI_BUILD}/lib/python${PYTHON_VERSION}:/home/ubuntu/jedi/ecflow-5.8.4/lib/python3.8/site-packages:${PYTHONPATH}"

5- Setup R2D2 (for MacOS and AWS Single Nodes)

If you are running skylab locally on the MacOS or an AWS single node instance, you will also have to setup R2D2. This step should be skipped if you are on any other supported platform. As with the previous step, it is recommended to complete these steps inside the python virtual environment that was activated above.

Clone the r2d2-data Repo

As with the other repositories, clone this inside your $JEDI_SRC directory.

cd $JEDI_SRC
git clone https://github.com/jcsda-internal/r2d2-data

Create a local copy of the R2D2 data store:

mkdir $HOME/r2d2-experiments-localhost
cp -R $JEDI_SRC/r2d2-data/r2d2-experiments-tutorial/* $HOME/r2d2-experiments-localhost

Install, Start, and Configure the MySQL Server

Execution of R2D2 on MacOS and AWS single nodes requires that MySQL is installed, started, and configured properly. For new site configurations see the spack-stack instructions for the needed prerequisites for macOS, Ubuntu, and Red Hat. Note, if you are reading these instructions, it is likely you have already setup the spack-stack environment.

You should have installed MySQL when you were setting up the spack-stack environment. To check this, enter brew list to the terminal and check the output for mysql.

Follow the directions for setting up the MySQL server found in the R2D2 tutorial starting at the Prerequisites for MacOS and AWS Single Nodes Only section. (If the link doesn’t work, the directions can be found in the TUTORIAL.md file in the r2d2 repository).

Note: The command used to setup the the local database should be run from the $JEDI_SRC/r2d2 directory. And the r2d2-experiments-tutorial.sql file is in $JEDI_SRC/r2d2-data.

6- Run SkyLab

Now you are ready to start an ecflow server and run an experiment. Make sure you are in your python virtual environment (venv).

To start the ecflow server:

ecflow_start.sh -p $ECF_PORT

Note: On Discover, users need to set ECF_PORT manually:

export ECF_PORT=2500
ecflow_start.sh -p $ECF_PORT

Please note “Host” and “Port Number” here. Also note that each user must use a unique port number (we recommend using a random number between 2500 and 9999)

To view the ecflow GUI:

ecflow_ui &

When opening the ecflow GUI flow for the first time you will need to add your server to the GUI. In the GUI click on “Servers” and then “Manage servers”. A new window will appear. Click on “Add server”. Here you need to add the Name, Host, and Port of your server. For “Host” and “Port” please refer to the last section of output from the previous step.

To stop the ecflow server:

ecflow_stop.sh -p $ECF_PORT

To start your ewok experiment:

create_experiment.py $JEDI_SRC/skylab/experiments/your-experiment.yaml

Note for MacOS Users:

If attempting to start the ecflow server on the MacOS gives you an error message like this:

Failed to connect to <machineName>:<PortNumber>. After 2 attempts. Is the server running ?

...

restart of server failed

You will need to edit your /etc/hosts file (which will require sudo access). Add the name of your machine on the localhost line. So if the name of your local machine is SATURN, then edit your /etc/hosts to:

##
# Host Database
#
# localhost is used to configure the loopback interface
# when the system is booting. Do not change this entry.
##
127.0.0.1     localhost SATURN
255.255.255.255       broadcasthost
::1       localhost

7- Existing SkyLab experiments

At the moment there are four SkyLab flagship experiments:

  • skylab-aero.yaml

  • skylab-atm-land.yaml

  • skylab-marine.yaml

  • skylab-trace-gas.yaml

To read a more in depth description of the parameters available and the setup for these experiments, please read our page on the SkyLab experiments description.