This is the part of the client where you can check the status of your submitted jobs and retrieve the results.
Once a job is submitted, the UNICORE NJS server tracks its status, and by using the Job Monitoring functions of the user client, the status can be monitored, and job groups can be killed or put on hold. After a job is terminated (normally, by job monitor commands or because of a job failure), the user can retrieve the job results (standard output, standard error, result files) and transmit them to the local workstation. After all results have been retrieved, the job information can be purged from the UNICORE system.
A UNICORE job is actually run under control of a local batch system in one of the batch queues on the execution system. As far as the UNICORE system is concerned, the following job states (both for job groups and for tasks) are supported
SUCCESSFUL: job or task has been run successfully. For a job, this means that all components have run successfully.
FAILED: execution of the job group or task has failed. For a job, this means that at least one of the components has failed, while other components may have run successfully. The end-user can inspect the status of the components to find out which have failed. For a task, this means that the task has failed. A message indicating the reason of failure can be displayed.
PENDING: job or task is queued within the UNICORE system. This happens if a predecessor task has not yet been executed, or if the whole job has not yet been started at all.
QUEUED: job or task is queued in the target batch system. This happens if the resources on the target system are used already by other (non-UNICORE) jobs.
EXECUTING: job or task is currently executing, that is running in the target batch system and is not yet completed. For a job group, this is indicated if at least one component is being executed