This tutorial is intended to help users run the eAIRS RCP application. This guide will use the sample data and eAIRS CFD workflow analysis to introduce the user to the most common features of the software. Since the software is still in early development stages, we are only distributing zip files containing the required files. Before distributing release clients, binary installation files will be created to guide the user through installing the application and providing shortcuts for launching the software.
The eAIRS RCP application can run on any platform that is supported by Eclipse, which is all major platforms. The RCP application simply has to be built for the platform that the user intends to run on. The initial builds being offered are Windows, Linux and Mac OSX (32 bit and 64 bit). The main purpose of the RCP application is to provide an easy to use interface for initiating and managing HPC jobs so the system requirements are minimal. The next section will talk in more detail about the system requirements for the RMI service that manages the execution of jobs. Some of requirements for the RMI Service can and will be eliminated in future releases.
Although eAIRS RCP is designed to launch jobs on any defined HPC machine, the execution is currently limited to grid-abe.ncsa.teragrid.org since Abe has ELF installed on it. ELF is required for launching the HPC jobs for eAIRS RCP. Because Abe is the only machine with ELF, the user will need an account on Abe to use the software. This limitation is easily overcome if another machine has ELF installed on it. As part of the deliverable of this software, ELF will be deployed on machines at KISTI.
Another requirement is that the eAIRS CFD executable should be installed on Abe in your home directory. The current version of the workflow allows you to specify the path to the executable in the User Interface, as you will see later, so you can place the unzipped files anywhere you have permissions.
Find your operating system and click the download link to download eAIRS. To install eAIRS, simply unzip downloaded file somewhere on your machine. To launch the application, do the following:
For Windows and Linux:
For Mac:
When logging into the application, you can check the box to save your password for next time if you are choosing to log in. After making your selection, you should see the main application screen similar to the one below.
This section will provide a brief explanation of each view seen when the application first starts.
The eAIRS Scenarios view allows users to create new scenarios, add data to the scenario and execute workflows on the data contained in the scenario. To create a new scenario, do the following:
You can also use the new scenario icon ( ) to bring up the same menu. Well see more of this view later as we add data from the repository and execute a workflow.
This view is where you can access information about your running workflows by connecting to an RMI Service endpoint. We haven't defined any RMI Services (see Service Registry View) so we cannot connect to an RMI service. Once we have defined a service, you can connect to it by clicking on the
icon and selecting your RMI service. We'll go through this later after we have defined an RMI Service.
The Basic Service Configuration view and the Advanced Service Configuration view are two PTPFlow views that allow users to tweak the settings used to launch your own RMI Service. Most users will not need to work with these views because as part of the final release, we will be setting up a remote RMI Service that is pre-configured. For this tutorial, we assume that the user will use or test RMI Service called Vertex, a pre-configured RMI Service that users can connect to using a few simple settings. The rest of this section is for administrators who are interested in knowing more about these two views and what content they have.
For those who cannot use our test RMI Service (explained later), continue reading to learn about what settings you must tweak before launching your RMI Service. The settings we will need to change are under the *Advanced Service Configuration_ view. To open it, do the following:
The first change we need to make is to the file user.auth.properties. Locate the file the CONFIGURATION FILES panel and double click on it to open it in an editor. You will need to add your username to this file. For example, if your username is foo, add:
foo=ADMIN
"/C=US/O=National Center for Supercomputing Applications/CN=FirstName LastName" foo
"/C=US/O=National Center for Supercomputing Applications/OU=People/CN=FirstName LastName" foo
The first entry is used if the user with username "foo" logs in to the NCSA MyProxy using their Kerberos password; otherwise, the NCSA CA would be applicable (the one with OU=people). FirstName would be replaced with foo's real first name and LastName would be replaced with foo's last name. After making the changes, save the file by closing it or ctrl+s.
For now, we are assuming that this workflow will run on NCSA Abe so the rest of the default parameters should be ok. In the future, there will be a tutorial on how to edit more of this view when trying to add new machines.
The other service configuration view, Basic Service Configuration can be found below. This view allows users to change various parameters such as the MyProxy server URI, refresh time, etc. Most users will not need to make any changes to this view. For administrators who want to know more about this view, see PTPFlow's documentation.
The repository view shows all imported datasets that can be used in workflows that are available with the software (more on that in the developer documentation). Two sample datasets have been included for the eAIRS CFD Workflow and can be automatically imported by clicking on the triangle in the Repository view and selecting Install Sample Data. In the image below, the view menu (triangle) is being pointed at by the mouse cursor.
After clicking on Install Sample Data, the sample data will be installed in a few moments. Once it finishes, click on the + symbol next to dataset to see the two installed datasets. To add a dataset to the scenario we created, click on the dataset in the repository view and drag it to your scenario. Do this for both datasets that we installed and your scenarios view should look similar to the image below:
You could also use shift+click to select both datasets and drag them to your Scenario.
Note: These sample datasets are intended for the eAIRS executable that can produce VTK result files. They will not work with the other executable that does not produce VTK files.
This view is the advanced view for defining an RMI Service that will be used to launch jobs on a supercomputer. For the final release of the software, we will document and setup a standalone RMI Service for KISTI users so they can just point their client at the RMI Service. The details provided below are for setting up your own RMI Service to use.
The current RMI service is limited to Unix-based operating systems so the RMI service must be launched on a Linux machine, Mac or other Unix-based computer. For this tutorial, we will assume that the machine running the RCP application is Unix based and can run the RMI service. To add a service, click the icon in the view.
This view will display metadata associated with a dataset by clicking on a dataset in either the Repository view or the Scenario's view. Add a tag to a dataset and then click on it while viewing the Metadata view to see your new tag. For example, if you click on "eAIRS Input File" in the Repository view, your Metadata view should look similar to the one below:
The tag for this dataset will be used to verify that an input is valid for a field in the eAIRS analysis wizard.
This section discusses some of the preferences that can be set in the Preference menu of eAIRS.
For the eAIRS application, users can optionally set a preference that specifies the location of the eAIRS executable used in the workflow analyses. By setting this option, the analysis field will populate automatically from the preference. Users can still change the field to a different location, if desired. To set the preference, do the following:
For this tutorial example, we will assume that the user can use our standalone RMI Service. The only difference between this tutorial and the final release to KISTI will be the replacement of our RMI Service with their standalone RMI Service. So, the first step is to add our RMI Service at vertex.ncsa.uiuc.edu
Enter Vertex for the name of the service, rmi://vertex.ncsa.uiuc.edu:1099/ptpflow for RmiContact(URI), 51001,52000 for RmiPortRange and 50000,51000 for GridFTPPortRange. After entering this information, click Finish. You should see the name of the service next to the icon in your scenarios view.
Now that we have an RMI service, it's time to connect to it so we can launch our eAIRS CFD workflow.
Once the workflow is submitted, you will see a confirmation similar to the one below:
If you click the Refresh icon in the Workflow Overview view, you should see a view similar to the one below:
If you double click on your username (see the mouse cursor in the above view) or click the Show button, you should receive some information about the workflow similar to the image below:
If you right click on your workflow (see the mouse cursor position in the above image), you should have some options such as Graph Details, Monitor Workflow Events, etc. If you click on Monitor Workflow Events, a new view called Event Monitor should open. Click on the Connect icon and Click Finish to accept the default options. You should see some events similar to the ones below:
The PTPFlow views offer many options for obtaining various details about your workflow and more of these will be detailed later in other tutorials and in some short videos that will provide more information about how to access the various features of PTPFlow.
After your workflow finishes, make sure your scenario is selected in the Scenarios View and then go back to the Workflow Overview view and right click on your workflow job. You should see an option called ( ) Retrieve Results, select it. This will download the results for the workflow to your repository and add them to your Scenario. Each result file will be tagged "result". Depending on the speed of your connection, this might take a few minutes to complete.
If you have Paraview installed on your machine, go to your Scenarios View and do the following:
Paraview should open and you should see an image similar to the one below:
To view the results, press the "Apply" button and you should now see a screen like the one below:
This is similar to the Single run workflow, except the user can specify multiple input files and meshes.
Once the workflow is submitted, you will see a confirmation similar to the one below:
This section discusses how you can create your own input files to use with eAIRS. There are two options, you can either create your own input file outside of eAIRS RCP and import the file into the repository or you can use eAIRS RCP to help you create new input files. The next section discusses the latter.
To create a new input file with eAIRS RCP, do the following:
This should have created 3 new files in your Repository called MyInput_0, MyInput_1, MyInput_2. You might have to click the icon to update the Repository. To create a single file, follow the previous steps, but only specify initial values for each variable.
Note: be careful specifying too many parameterizations since the number of files can grow quite large. For example, using 3 different mach numbers, 10 different Reynolds numbers and 3 different Angles of Attack will generate 90 input files.
To tag a dataset, do the following:
To see your new tags, open the Metadata view and click on the dataset if it is not already selected.
To add a comment to a dataset, do the following: