APT-INSAR on RMACC Summit

This repository contains a bash script that will install The "apt-insar" Docker container maintained by the Alaska SAR Facility on Summit, such that it can be used with the Singularity containerization software.

---

Installation

To install the apt-insar docker container on Summit:

Step 1: Login to a Summit 'scompile' node and download this github repository to any directory you want, as follows:

git clone https://github.com/ResearchComputing/asf-insar-singularity.git

Step 2: Run the installation script:

cd asf-insar-singularity 
bash ./apt-insar-install.sh
cd ../ 
rmdir asf-insar-singularity

Note: The container will be installed in /projects/$USER/containers by default. You can change the INSTALL_DIR variable in the script if you want it somewhere else.

Step 3: Ensure everything was installed:

cd /projects/$USER/containers
ls

You should see four files: apt-insar apt-insar.sif commands.sh run-apt-insar-job.sh

• apt-insar: the "wrapper script" that you invoke to run the container
• apt-insar.sif: the container (sif="Singularity Image Format")
• commands.sh: the script invoked by the container when you run apt-insar
• run-apt-insar-job.sh: an example job script for running the container on Summit (additional details below).

---

Usage

To use the ASF apt-insar Docker container on Summit, use this command in your job script:

/projects/$USER/containers/apt-insar --reference-granule REFERENCE_GRANULE --secondary-granule SECONDARY_GRANULE [--username USERNAME] [--password PASSWORD] [--dem {ASF,SRTM}]

| Option | Description |

| --reference-granule | Reference granule name. |

| --secondary-granule | Secondary granule name. |

| --username | Earthdata Login username. |

| --password | Earthdata Login password. |

| --dem | Digital Elevation Model. ASF automatically selects the best geoid-corrected NED/SRTM DEM. SRTM uses the ISCE default settings. |

For example [note that the "\" merges lines]:

/projects/$USER/containers/apt-insar \
--reference-granule S1A_IW_SLC__1SDV_20190716T135159_20190716T135226_028143_032DC3_512B \
--secondary-granule S1A_IW_SLC__1SDV_20190704T135158_20190704T135225_027968_032877_1C4D \
--username $EARTHDATA_USER \
--password $EARTHDATA_PASS

Note in the example above it is assumed that you have saved your username and password to environment variables called $EARTHDATA_USER and $EARTHDATA_PASS. One way to ensure that these variables persist across all of your sessions is to export them at the start of any session by placing the following two lines in ~/.bash_profile (e.g, for user=janedoe and password=1Lov3NASA!):

export EARTHDATA_USER=janedoe
export EARTHDATA_PASS=1Lov3NASA!

A sample job script called run-apt-insar-job.sh can be found in your /projects/$USER/containers directory. You can edit this script to suit your workflow and submit jobs to Summit as follows: sbatch run-apt-insar-job.sh

Note that the job script is presently configured to use one core (--ntasks=1). Each core on the Summit shas partition is associated with 4.48 GB RAM. If your job crashes because it has insufficient memory, you can increase the number of cores (e.g., setting --ntasks=2 will give you 9.96 GB of RAM).