11.32. pegasus-status
1 pegasus-status Pegasus workflow- and run-time status
pegasus-status [-h|--help] [-l|--long] [-d|--debug] [-w|--watch <sec>] [-L|--legend] [--noqueue] [-j|--jsonrv] [rundir]
11.32.1. Description
pegasus-status shows the current state of the Condor Q and a workflow, depending on settings. If no valid run directory could be determined, including the current directory, pegasus-status will show all jobs of the current user and no workflows. If a run directory was specified, or the current directory is a valid run directory, status about the workflow will also be shown.
Many options will modify the behavior of this program, such as : watch mode, the presence of jobs in the queue, progress in the workflow directory, etc. It’s important to note that -j option takes precedence over all other options and will result in returning the status in a json structure format.
11.32.2. Options
- -h; –help
Prints a concise help and exits.
- -V; –version
Prints the version information and exits.
- -w [sec]; –watch [sec]
This option enables the watch mode. In watch mode, the program repeatedly polls the status sources and shows them in an updating window. This option requires an argument sec to determine how often these sources are polled.
We strongly recommend to set this interval not too low, as frequent polling will degrade the scheduler performance and increase the host load. The sec argument recommended to be entered is 60 seconds.
Watch mode is disabled by default.
- -L; –legend
This option shows a legend explaining the columns in the output
By default, legends are turned off to save terminal real estate.
- -Q; –noqueue
This option turns off the output from parsing Condor Q.
By default, Condor Q will be parsed for jobs belonging to the workflow run directory specified as an arguemnt.
- -d; –debug
This is an internal debugging tool and should not be used outside the development team. In debugging mode, it will show the condor Q expression used to retrieve the jobs in Q belonging to the workflow run directory entered. It also shows the traversal of all the DAG sub-directories’ used to retrieve the progress of the DAG (and sub-DAGs, if any).
By default, debug mode is off.
- -l; –long
This option will show one line per sub-DAG, including one line for the workflow. In case of hierarchical workflows, now each sub-dag is shown along with indented branch. This option now shows more columns for jobs in condor Q.
By default, only DAG totals (sums) are shown.
- -D; –dirs
This option can be used along with –long option in case of hierarchical workflows, to show relative directories under dagnames per sub-DAG. It’s main usage comes for hierarichal workflows.
By default, only dag names are shown.
- -j
This option returns the status of the workflow in a JSON serializable data structure (Python dict). Sample of this structure is shown below, where the keys are -
totals : contains the overall progress of the workflow
dags : contains progress regarding each workflow, in case of hierarchical workflows each sub-DAG with it’s name as corresponding key
condor_jobs: contains all the jobs in Q belonging to a specific workflow, with it’s unique wf_uuid as corresponding key. Furthermore, each workflow has DAG_NAME key and DAG_CONDOR_JOBS key with a list of condor Q jobs of the corresponding DAG
{
"totals": {
"unready": 13,
"ready": 0,
"pre": 0,
"queued": 1,
"post": 1,
"succeeded": 14,
"failed": 0,
"percent_done": 48.28,
"total": 29
},
"dags": {
"root": {
"unready": 4,
"ready": 0,
"pre": 0,
"queued": 1,
"post": 0,
"succeeded": 8,
"failed": 0,
"percent_done": 61.54,
"total": 13,
"dagname": "workflow-0.dag",
"state": "Running"
}
},
"condor_jobs": {
"f436c93a-5ef5-4d9f-815a-0ccee5e9de67": {
"DAG_NAME": "root",
"DAG_CONDOR_JOBS": [
{
"ClusterId": 2457,
"Cmd": "/usr/bin/pegasus-dagman",
"EnteredCurrentStatus": 1664819625,
"Iwd": "workflow/run/directory/run0001",
"JobPrio": 0,
"JobStatus": "Run",
"pegasus_site": "local",
"pegasus_wf_name": "hierarchical-workflow-0",
"pegasus_wf_xformation": "pegasus::dagman",
"UserLog": "run/directory/run0001/workflow-0.dag.dagman.log"
},
{
"ClusterId": 2465,
"Cmd": "/usr/bin/condor_dagman",
"EnteredCurrentStatus": 1664819681,
"Iwd": "workflow/run/directory/run0001/00/00/./inner",
"JobPrio": 30,
"JobStatus": "Run",
"pegasus_site": "local",
"pegasus_wf_dag_job_id": "pegasus-plan_diamond_subworkflow",
"pegasus_wf_dax_job_id": "diamond_subworkflow",
"pegasus_wf_name": "hierarchical-workflow-0",
"pegasus_wf_xformation": "condor::dagman",
"UserLog": "run/directory/run0001/workflow-0.log"
}
]
}
}
}
- rundir
This option shows statistics about the given DAG that runs in rundir. To gather proper statistics, pegasus-status needs to traverse the directory and all sub-directories. This can become an expensive operation on shared filesystems.
If the current directory is not a valid rundir, no DAG statistics will be shown.
11.32.3. Return Value
pegasus-status will typically return success in regular mode, and the termination signal in watch mode. Abnormal behavior will result in a non-zero exit code.
11.32.4. Example
- pegasus-status rundir
This invocation will parse the Condor Q for the current user and show all her jobs. Additionally, if the current directory is a valid Pegasus workflow directory, totals about the DAG in that directory are displayed.
- pegasus-status -l rundir
As above, but providing a specific Pegasus workflow directory in argument rundir and requesting to itemize sub-DAGs.
- pegasus-status -w 300 -Ll
This invocation will parse the queue,show legends, itemize DAG statistics of the current working directory, and redraw the terminal every five minutes with updated statistics.
11.32.5. Restrictions
Currently only supports a single run directory. If you want to watch multiple run directories, we suggest to open multiple terminals and watch them separately. If that is not an option, or deemed too expensive, you can ask pegasus-support at isi dot edu to extend the program.
11.32.6. See Also
condor_q(1), pegasus-statistics(1)