CONDOR_TAIL(1)

✨ condor_tail ✨

okay bestie, this is condor_tail, from the HTCondor Manual — it's the lil command that lets you peek at the last bits of a running job's standard output (or any file in its sandbox) without waiting for the whole thing to finish 💖

---

💫 NAME 💫

condor_tail — display the last contents of a running job's standard output or file. no cap, that's literally the whole job description 🫶

---

🌸 SYNOPSIS 🌸

condor_tail [-help | -version]

condor_tail [-pool hostname[:portnumber]] [-name scheddname] [-debug]
[-maxbytes numbytes] [-auto-retry] [-follow] [-no-stdout] [-stderr]
job-ID [filename ...]

---

🦋 DESCRIPTION 🦋

So condor_tail shows you the last bytes of a file sitting in the sandbox of a running job — and you tell it which job by passing job-ID on the command line. By default it tails stdout (the main character, obviously 👑).

The number of bytes it shows is capped at 1024, unless you go change it with the -maxbytes option. And heads up: that limit applies to each individual tail of a file. Like, if you're following a file, the limit gets applied every single subsequent time output comes in. so it's a per-tail cap, not a per-session cap 🎀

And if you pass a filename, that name has to be explicitly listed in the job's transfer_output_files submit command. no sneaking in random files, bestie, them's the rules 🙈

---

✨ OPTIONS ✨

• -help — shows you the usage info and bounces. that's it, that's the whole vibe.
• -version — spills the version info then exits.
• -pool hostname[:portnumber] — picks a pool by handing over the central manager's host name, plus an optional port number if you're feeling specific 💅
• -name scheddname — queries the condor_schedd daemon identified by scheddname.
• -debug — gives you the extra debugging tea ☕
• -maxbytes numbytes — caps the max number of bytes transferred per tail access. don't specify it? then the max is 1024.
• -auto-retry — if the job isn't running yet, it retries that tail every 2 seconds. persistent queen ⭐
• -follow — tails the file(s) over and over, until you interrupt it.
• -no-stdout — don't tail stdout. simply not doing it.
• -stderr — tails stderr instead of stdout.
• job-ID — the job identifier, in the format cluster.proc, or just cluster if you want to tail the first job in the cluster.
• filename — optional filename(s) to tail instead of stdout. Must be listed in transfer_output_files (we said it once, we'll say it again, it's a hard requirement fr fr).

---

🌈 GENERAL REMARKS 🌈

condor_tail is iconic for monitoring how a running job is progressing without sitting around waiting for it to complete and transfer its output files back. instant gratification, we love her 💜

And the -follow option behaves a lot like tail -f does on Unix systems — it just keeps showing new output as the job writes it. like a livestream but for your job's output 🍓

---

💖 EXAMPLES 💖

Tail stdout of job 123.0:

$ condor_tail 123.0

Tail stderr of job 123.0:

$ condor_tail -stderr 123.0

Follow stdout of job 123.0, updating every 2 seconds:

$ condor_tail -follow 123.0

Tail a specific output file from job 123.0:

$ condor_tail 123.0 output.log

Tail with an increased byte limit:

$ condor_tail -maxbytes 4096 123.0

---

🎀 EXIT STATUS 🎀

• 0 — Success 💚
• 1 — Failure 💀

---

🧚 SEE ALSO 🧚

condor_submit, condor_q, condor_ssh_to_job, condor_transfer_data

---

⭐ AVAILABILITY ⭐

Linux, MacOS, Windows — she's everywhere, she's universal 🌍

---

👑 AUTHOR 👑

HTCondor Team

---

💜 COPYRIGHT 💜

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!! this came from the HTCondor Manual and it's dated Oct 16, 2026 — go tail something and slay 💅✨🦋