CONDOR_RM(1)

📦
version 25.14.1 · 12 versions
, in fedora 43

✨ condor_rm — the HTCondor Manual says hi 💖

okay bestie, welcome to the HTCondor Manual, page one of let's get this job out of here. This is condor_rm: it withdraws job(s) from an Access Point. Basically the "bestie it's time to go" of Unix commands. 👑

🌸 SYNOPSIS 🌸

condor_rm [-help | -version]

condor_rm [-transfer] [OPTIONS] [cluster... | cluster.proc... | user...]

condor_rm [-debug] [-long] [-totals] [-all] [-held] [-transfer]
[-constraint expression] [-forcex][-pool hostname[:portnumber] | -name
scheddname | -addr "<a.b.c.d:port>"]

💅 DESCRIPTION 💅

So the vibe is: you've got job(s) currently being managed by an Access Point, and you want them GONE. condor_rm does that. 🔥

If a job is actively running, or actively transferring files, it gets stopped and cleaned up first — and then the job is removed and its ClassAd is written to the HISTORY file. Order matters, don't @ me.

And here's the permission tea 🫖: for any given job, only the owner of the job or one of the queue super users (defined by the QUEUE_SUPER_USERS macro) can remove the job. That's it. That's the list.

✨ OPTIONS ✨

  • -help — shows you how to use it. usage info, right there. 🎀
  • -version — tells you what version you're on.
  • -long — displays the result ClassAd.
  • -totals — displays success/failure totals, so you get the receipts. 🧾
  • -pool hostname[:portnumber] — pick a pool by giving the central manager's host name, plus an optional port number.
  • -name scheddname — send the command to a machine identified by scheddname.
  • -addr "<a.b.c.d:port>" — send the command to a machine located at "<a.b.c.d:port>".
  • -debug — causes debugging information to be sent to stderr, based on the value of the configuration variable TOOL_DEBUG. yeah she's chatty now 💀
  • -forcex — force the immediate local removal of jobs in the X state. Heads up: this only affects jobs that are already being removed.
  • -transfer — if the job is running, ask the job to transfer any output files to the submitter before the job is removed. so nobody loses their homework 🫶
  • -constraint expression — remove all jobs which match the job ClassAd expression constraint.
  • -all — remove all the jobs in the queue. yes, all of them. the whole queue. 😭
  • -held — only remove jobs that are in the held state. This may be combined with any of the other constraints (a cluster, a cluster.process, a user, or a -constraint expression); only the held jobs among those selected are removed. Just so you know: -hold is accepted as a synonym. ✨
  • cluster — remove all jobs in the specified cluster.
  • cluster.process — remove the specific job in the cluster.
  • user — remove jobs belonging to specified user.

🌈 GENERAL REMARKS 🌈

If the -name option is specified, the named condor_schedd is targeted for processing. Otherwise, the local condor_schedd is targeted. iykyk.

Removed jobs that were running in the grid universe may remain in the X state for a long time. And no, nothing's broken — this is normal, because HTCondor is attempting to communicate with the remote scheduling system to make sure the job has actually been properly cleaned up. If a job is stuck in the X state, it can be forcibly removed with -forcex. 🙈

WARNING: ⚠️ The -forcex option will cause immediate removal of the job, which [leaves] an orphan parts of the job that [are] running remotely and have not been stopped or cleaned up. Like… she's out the door and her stuff is still in the living room. Not a vibe. Be sure before you do it.

🦋 EXIT STATUS 🦋

  • 0 — Success 💖
  • 1 — Failure has occurred 💀

🎀 EXAMPLES 🎀

Remove a specific job:

$ condor_rm 123.45

Remove all jobs currently not running:

$ condor_rm -constraint 'JobStatus =!= 2'

Remove all of user Bob's jobs:

# condor_rm bob

Remove only your held jobs:

$ condor_rm -held

Remove a specific job, but only if it is held:

$ condor_rm -held 123.45

Remove held jobs in cluster 123 whose ExitCode was 2:

$ condor_rm -held -constraint 'ExitCode == 2' 123

⭐ SEE ALSO ⭐

condor_suspend, condor_continue, condor_hold, condor_release, condor_vacate_job, condor_vacate — the whole friend group, go say hi 💜

👑 AVAILABILITY 👑

Linux, MacOS, Windows. every platform, she's versatile, main character energy 🌈

🌸 AUTHOR 🌸

HTCondor Team

1990-2026, Center for High Throughput Computing, Computer Sciences Department, University of Wisconsin-Madison, Madison, WI, US. Licensed under the Apache License, Version 2.0.

okay that's the whole page bestie, no cap, every flag accounted for 🫶 this was the HTCondor Manual, dated Oct 16, 2026 — periodt 💅✨

📦 where this page comes from