---

THE JUST-THE-BASICS GUIDE TO GETTING STARTED WITH SCISERVER

---

NASA and Johns Hopkins University have developed SciServer, which allows researchers to reprocess their data with SAS remotely on servers at JHU. People who are interested are strongly encouraged to take their time and examine the tutorials and documentation available on the SciServer site in order to best take advantage of this powerful new system. However, the GOF acknowledges that there are times when all a person really wants is a fast and simple on-ramp. This is intended to be just that: a quick-start guide to using SAS in SciServer.

Before getting started, it is important to know that SciServer, while very powerful, cannot handle visualizations. Therefore, users must have visualization software (e.g. ds9 and fv) on their own computers and download the files they want to see. Similarly, if a person wants to fit a spectrum, they would need to have fitting software (e.g. Xspec or Sherpa) on their own computer and download the spectrum and response files from SciServer.

Now on with the show!

The first thing you will need to do is make an account and log in:

Once you have logged in, a new tab will open that shows the SciServer Dashboard (see Figure 2). On the Dashboard, click on Compute:

This will open another tab that shows the SciServer Compute - Interactive Work page (see Figure 3), which will give you the opportunity to start a container. Before doing so, it is strongly recommended that you read the information about file storage. There are three main takeaways:

1. if you put data or scripts on the server that you would like to see again, make sure they are kept in either the Storage or Temporary directory
2. files stored in the Storage and Temporary directories persist between containers, but only those in Storage are backed up
3. the Temporary directory is 40T (shared between users), while the Storage directory is 10G (per user), so it is recommended that people use Temporary as their workspace and copy whatever they want to archive to Storage.

With that understanding, we can now click on "Create Container".

This will open a pop-up window (see Figure 4) where you can name the container and define its attributes in the Compute Image drop-down menu, and User Volumes and Data Volumes selections:

For SAS, we need to select "HEASARCv6.29" in Compute Image. We also want a persistent Storage Volume and scratch Temporary Volume in User Volumes. Under Data Volumes, we will want "HEASARC data" and "HEASARC software". With these set, we can now create the container:

Now in the Interactive Work tab, you will see your container and information about when it was created, its domain, compute image, and status (see Figure 6). There are also small icons that let you stop a running container, get information about it, or delete it.

Clicking on the name of the container will open a JupyterLab tab (see Figure 7), which has a launcher to let you choose an interface. In the JupyterLab tab, select "Terminal".

This will open a traditional linux terminal in the main window of the launcher tab, and a smaller window to its left with graphical directory navigation. Be aware that these windows are not linked -- moving to a different directory in one window does not move you to that directory in the other.

Taking a closer look at the navigation window, above the box that says "Filter files by name", there are four icons (see Figure 9). These let you open a new launcher window, make a new folder, refresh the file list, and upload files. Double-clicking on a folder name will open it; right-clicking on a file will open a menu of options for that file, one of which is to download it.

Now that we have a terminal, we can get some data and initialize SAS. If we are working on Lockman Hole data (Obs ID 0123700101), we can set up the workspace and final repository like this:

mkdir workspace/Temporary/[your_username]/scratch/LockmanHole

mkdir workspace/Storage/[your_username]/persistent/LockmanHole

SciServer has direct access to archived data, and we know which Obs ID we want, so we can copy it over easily and uncompress it.

cd workspace/Temporary/[your_username]/scratch/LockmanHole

mkdir -p 0123700101/ODF

cp /FTP/xmm/data/rev0/0123700101/ODF/* 0123700101/ODF

cd 0123700101/ODF

gunzip *gz

Now we need to initialize SAS and set the standard environment variables:

source /opt/xmmsas/xmmsas_20210317_1624/setsas.sh

export SAS_CCFPATH="/FTP/caldb/data/xmm/ccf"

export SAS_ODF=`pwd`

export SAS_ODFPATH=`pwd`

And now we can run the SAS setup tasks and reset the environment variables as usual, in preparation to run the standard processing tasks:

cifbuild

export SAS_CCF=`pwd`"/ccf.cif"

odfingest

export SAS_ODF=`pwd`'/'`ls *SUM.SAS`

cd ..

mkdir PROC

cd PROC

At this point, the standard repipelining software (emchain or emproc; epchain or epproc) can be run and the usual filtering can be done as shown in the ABC Guide, keeping in mind the caveat that light curves, event files, etc. must be downloaded to your local computer in order to view them. When you are done, remember to copy things over to Storage. To copy only the processed data, type

cd /home/idies/workspace/Temporary/[your_username]/scratch/LockmanHole/0123700101

mkdir /home/idies/workspace/Storage/[your_username]/persistent/LockmanHole/0123700101

cp PROC /home/idies/workspace/Storage/[your_username]/persistent/LockmanHole/0123700101

or, if you want to make a tarball of the ODF and reprocessed directories and move it to Storage, type

cd /home/idies/workspace/Temporary/[your_username]/scratch/LockmanHole/

tar cvfz 0123700101.tgz 0123700101/*

mv 0123700101.tgz /home/idies/workspace/Storage/[your_username]/persistent/LockmanHole

It is also strongly recommended that you download your data to your local machine for archiving.

Lastly, it should be noted that, upon occasion, SciServer gets overloaded. When this happens, users cannot access their containers. In that case, it is recommended that users try again a bit later when the server is not as busy. As a last resort, users might need to remake their containers (i.e., delete their current container and rebuild it from the HEASARC image.) In this case, anything that was not stored in Temporary or Storage is deleted.

---

If you have any questions concerning XMM-Newton send e-mail to xmmhelp@lists.nasa.gov

This file was last modified on Wednesday, 27-Jan-2021 14:44:31 EST
Curator:Michael Arida (ADNET); michael.arida@nasa.gov