Quick reference: HTCondor commands
Introduction
This page lists common HTCondor commands and options for jobs. Users familiar with HTCondor and job submission on CHTC's High Throughput Computing (HTC) system can use this page as a quick reference. For users who are just starting out, we suggest reading the full guides (linked) to understand the full context of each command or option.
Please note the following!
- Bracketed items (
<>) denote where to place your input. Do not include the brackets in your command.- All commands should be entered on a single line.
Submit jobs
You can use these commands to submit, hold, or remove your jobs.
| Command | Use | Notes |
|---|---|---|
condor_submit <submit_file> | submits job(s) as specified by submit_file | See job submission |
condor_submit -i <submit_file> | submits an interactive job as specified by submit_file | |
condor_rm <username> | removes all of your jobs | |
condor_rm <job_ID> | removes the job(s) associated with job_ID | |
condor_hold <job_ID> | holds the job(s) associated with job_ID | |
condor_release <job_ID> | releases the held job(s) associated with job_ID | |
condor_ssh_to_job <job_ID> | allows you to "ssh" to the execution point on which the job associated with job_ID is running on |
Monitor jobs
| Command | Use | Notes |
|---|---|---|
condor_q | displays status of your submitted jobs; jobs are batched by default | See monitor your jobs |
condor_q -nobatch | displays status of your submitted jobs without the batched view | |
condor_q <job_ID> | displays status of the job(s) associated with job_ID | |
condor_q -l <job_ID> | lists all attributes of the job(s) associated with job_ID | |
condor_q -hold <job_ID> | displays the hold reason for job(s) associated with job_ID | |
condor_q -better-analyze <job_ID> | displays simulated results of the matching process associated with the job | This is a starting point for troubleshooting jobs sitting in the idle state. |
condor_watch_q | displays the "real-time" status of your jobs | Updated every 2 seconds. CTRL + C to exit. |
condor_tail <job_ID> | prints the last 10 lines of the standard output the job associated with job_ID |
Machine information
These commands display information about the execution points - machines that execute/run jobs.
| Command | Use | Notes |
|---|---|---|
condor_status | lists all execution point slots | |
condor_status <execution_point> | lists information about the specified execution_point | |
condor_status -l <execution_point> | lists all attributes of execution_point | |
condor_status -compact | lists all execution_point slots in a compact view | |
condor_status -gpus | lists all execution_point slots with GPUs |
Glossary
| Term | Meaning |
|---|---|
| access point | The machine which you log into to access CHTC's servers. This is the machine you use to prepare files for job submission and submit jobs. |
| cluster ID | A unique number associated with a single job submission. |
| error file / standard error | The file in which your job typically prints any errors. |
| execution point | The machine that executes or runs your job. |
| held/hold | A job state in which your job has stopped running due to an error. |
| idle | A job state in which your job has not yet matched to an execution point and hasn't started running yet. |
| job ID | The unique number associated with a job. This consists of a cluster ID, followed by a period and a process ID. For example, the job ID 12345.0 has a cluster ID of 12345 and a process ID of 0. |
| log file | The file where HTCondor prints messages about your job's execution and resource usage. |
| output file / standard out | The file where your job typically prints output. Any messages printed to the "screen" in a job will be saved in this file. |
| process ID | A unique number associated with each job within a job submission. See submit multiple jobs. |
| running | A job state in which your job has matched to the execution point and is currently executing/running. |
| submit file | A text-based file that tells HTCondor details about your job, including the commands to run, what files need to be transferred, where to save the outputs, and more. See the job submission. |