The Elastic High Performance Computing (E-HPC) console provides job-related features that include creating or stopping a job and viewing the job status. This article explains how to easily manage jobs.
Note: Create an E-HPC cluster first, if you have not created one yet.
You have uploaded the job execution script and job input data to the Network Attached Storage (NAS) instance and created a user in the corresponding cluster before creating a job. For more information, see Submit a job.
Log on to the E-HPC console and click Job from the left-side navigation pane to visit the Job overview page.
Click Details to view the job details. You can stop a job in the “Queue” or “Running” status.
Click Create Job in the upper-right corner of the Job page.
In the dialog box, enter all the required job information.
Click OK to submit the job to the cluster for execution.
Job Name: The name of the job. This job name is displayed in the job list. If you want to automatically download a job file, the job name is used as a decompressed directory name.
Job Template: A job template is a pre-configured template, which helps to create the job quickly.
Username and Password: The username and password used for executing the job. You must create the user on the Users page in advance.
Command line: The command for executing the job submitted to the scheduler. It can either be a script file or a command text.
Post-processing command: The post-processing script after the job is executed.
Enable Job Array: Specifies whether to enable the job array function of the scheduler.
Queue: If the compute node has joined the specified QUEUE, the job needs to be submitted to the corresponding specified Queue; if not specified, the job will be submitted to the scheduler’s default queue.
Standard Output Path and Error Output Path: Specify the stderr/stdout redirection paths (in the Linux operating system), which contain the output file names.
Environment variable: The environment variables required for executing the job. The environment variables are output to the job execution script and can be referenced in the script. Click a plus (+) sign to configure the environment variables by the name-value pair.
Edit Job File: Create, edit, and save job files online. When there are multiple job files, you can view, edit, and delete them by displaying the file list.
Use OSS Job File:
Job File: You can select a job file that has been uploaded to the OSS bucket. You can compress the job execution script, job input data file, and post-processing script into a package and upload it to the OSS bucket. E-HPC automatically downloads the package while executing the job.
Click Select File to open the OSS resource manager dialog box. The OSS buckets in the current area are displayed. If you have not created any OSS bucket, you can click the Upload Through OSS Console link in the resource manager dialog box and create an OSS bucket on the OSS console page. For more information about how to upload files to an OSS bucket, see OSS help documentation.
Note: To use this function activate an OSS bucket first.
- Decompress: After this option is enabled, E-HPC automatically decompresses the job file package before executing the job. Currently, job files in the ZIP, TAR, and GZIP formats are supported.
Note: E-HPC decompresses the package to the folder named the job name in the current working directory (your home directory). Therefore, to execute the scripts in the package, add the job name directory.