# Leftover files in /tmp

**URL:** https://discourse.openondemand.org/t/leftover-files-in-tmp/4718
**Category:** Get Help
**Created:** [January 21, 2026, 5:22pm UTC](https://discourse.openondemand.org/t/leftover-files-in-tmp/4718 "2026-01-21T17:22:33Z")
**Posts on this page:** 5
**Page:** 1

<div class="post-metadata">

### Author: ![davide-q](https://sea1.discourse-cdn.com/flex015/user_avatar/discourse.openondemand.org/davide-q/32/1510_2.png) [@davide-q](https://discourse.openondemand.org/u/davide-q)
#### Post date: [January 21, 2026, 5:22pm UTC](https://discourse.openondemand.org/t/leftover-files-in-tmp/4718/1 "2026-01-21T17:22:33Z")

</div>

I noticed that the simple VNC application which I installed on my cluster closely following the official documentation, leaves many files in `/tmp/dbus-*` and sometimes even in `/tmp/X1*-lock`

The former don’t seem to create any problem (they are even zero-sized, even though one could run out of inodes), but the latter prevent further jobs on the same node from starting correctly. The problem is occasional (I have not been able to track it down), and I have seen the related discussion at [Lock files for VNC-based apps not being cleaned up](https://discourse.openondemand.org/t/lock-files-for-vnc-based-apps-not-being-cleaned-up/813)

Is there a best practice to prevent these files from accumulating? I know I can use the slurm epilogue and/or the OOD cleaner, but I’m looking for advice of what is the best approach for this and perhaps for a nudge to include such advice in the documentation (or make it more prominent, if it’s already there and I overlooked it).

Thanks

---

<div class="post-metadata">

### Author: ![maflister](https://sea1.discourse-cdn.com/flex015/user_avatar/discourse.openondemand.org/maflister/32/139_2.png) [@maflister](https://discourse.openondemand.org/u/maflister)
#### Post date: [January 22, 2026, 2:37pm UTC](https://discourse.openondemand.org/t/leftover-files-in-tmp/4718/2 "2026-01-22T14:37:03Z")

</div>

If you use SLURM, you can try [Slurm Workload Manager - job\_container.conf](https://slurm.schedmd.com/job_container.conf.html) . That is basically polyinstantiation on the compute nodes for temp directories and would likely solve this problem for any job, not just OOD.

---

<div class="post-metadata">

### Author: ![jeff.ohrstrom](https://sea1.discourse-cdn.com/flex015/user_avatar/discourse.openondemand.org/jeff.ohrstrom/32/136_2.png) [@jeff.ohrstrom](https://discourse.openondemand.org/u/jeff.ohrstrom)
#### Post date: [January 22, 2026, 3:15pm UTC](https://discourse.openondemand.org/t/leftover-files-in-tmp/4718/3 "2026-01-22T15:15:17Z")

</div>

Thank you @maflister! Yes that’s the same advice I would have given. It’ll give every job their own tmpfs tmp directory that disappears after the job completes.

---

<div class="post-metadata">

### Author: ![davide-q](https://sea1.discourse-cdn.com/flex015/user_avatar/discourse.openondemand.org/davide-q/32/1510_2.png) [@davide-q](https://discourse.openondemand.org/u/davide-q)
#### Post date: [January 22, 2026, 6:36pm UTC](https://discourse.openondemand.org/t/leftover-files-in-tmp/4718/4 "2026-01-22T18:36:58Z")

</div>

Thanks so much folks!

---

<div class="post-metadata">

### Author: ![system](https://us1.discourse-cdn.com/flex015/uploads/osc/original/2X/b/bae70bd0ed39a3ae769c2108155f4cb3e9da8385.png) [@system](https://discourse.openondemand.org/u/system)
#### Post date: [July 21, 2026, 6:37pm UTC](https://discourse.openondemand.org/t/leftover-files-in-tmp/4718/5 "2026-07-21T18:37:02Z")

</div>

This topic was automatically closed 180 days after the last reply. New replies are no longer allowed.
