TMPDIR cleanup and Docs #2

marcnnn · 2024-11-26T16:59:44Z

No description provided.

josevalim · 2024-11-26T20:29:13Z

lib/flame_slurm_backend/slurm_client.ex

+    # remove the TMP folder
+    # this should be called since SLURM sends
+    IO.puts(file, """
+rm -rf $TMPDIR/$SLURMJOBID


We could set the tmpdir and remove it in the Elixir script itself. Would that be preferrable?

README.md

josevalim · 2024-11-26T20:35:51Z

README.md

+# Cleanup
+
+On some clusters TMPDIR is cleaned per Job on others not.
+This is why TMPDIR is changed to one based on the Job ID and deleted after SIGUSR1 is send by SLURM this is Configured by default in Flame SLurm to 30 Seconds before kill.


Suggested change

This is why TMPDIR is changed to one based on the Job ID and deleted after SIGUSR1 is send by SLURM this is Configured by default in Flame SLurm to 30 Seconds before kill.

This is why `TMPDIR` is changed to one based on the Job ID, which is automatically cleared when the Job terminates. More precisely, Slurm is automatically configured to send a SIGUSR1 signal to FLAME 30 seconds before it terminates the Job, which then deletes the directory.

README.md

lib/flame_slurm_backend/slurm_client.ex

Co-authored-by: José Valim <[email protected]>

TMPDIR cleanup and Docs

13b4e31