Skip to content

Conversation

@dmwnz
Copy link

@dmwnz dmwnz commented Nov 27, 2025

Related to snakemake/snakemake-executor-plugin-slurm#379 and matching PR snakemake/snakemake-executor-plugin-slurm#380

Summary by CodeRabbit

  • Bug Fixes

    • Improved script execution handling for SMP jobs by implementing stdin-based script streaming to the scheduler, replacing inline command embedding.
  • Improvements

    • Enhanced debugging capabilities with script content logging for better troubleshooting.

✏️ Tip: You can customize this high-level summary in your review settings.

@coderabbitai
Copy link

coderabbitai bot commented Nov 27, 2025

📝 Walkthrough

Walkthrough

Refactors job script invocation for SMP jobs to stream bash scripts via stdin instead of embedding them in command lines. Creates a local variable to hold the srun bash script, appends "bash -s" to the srun call, and passes the script through subprocess stdin in text mode.

Changes

Cohort / File(s) Summary
SMP job script streaming
snakemake_executor_plugin_slurm_jobstep/__init__.py
Introduces srun_script variable to hold job script; routes script to stdin using "bash -s" instead of inlining; enables text mode for subprocess; writes script to stdin if available; adds logging for debugging.

Sequence Diagram(s)

sequenceDiagram
    participant Code
    participant Subprocess
    participant Bash as Bash Process

    rect rgba(100, 150, 200, 0.1)
    note over Code: New Flow - Stdin-based Script Streaming
    Code->>Code: Format job script into srun_script
    Code->>Code: Append "bash -s" to srun call
    Code->>Subprocess: Create subprocess with stdin=PIPE, text=True
    Subprocess->>Bash: Launch bash -s
    Code->>Bash: Write srun_script to stdin
    Code->>Bash: Close stdin
    Bash->>Bash: Execute script from stdin
    Bash-->>Subprocess: Return exit code
    end
Loading

Estimated code review effort

🎯 2 (Simple) | ⏱️ ~10 minutes

  • Area of focus: Verify stdin handling correctness, edge cases when script is unavailable, and subprocess text mode compatibility with the bash script encoding.
  • Consideration: Confirm that closing stdin after script write doesn't prematurely terminate the bash process and review logging output for sensitive data leakage.

Possibly related issues

Pre-merge checks and finishing touches

❌ Failed checks (1 warning)
Check name Status Explanation Resolution
Docstring Coverage ⚠️ Warning Docstring coverage is 0.00% which is insufficient. The required threshold is 80.00%. You can run @coderabbitai generate docstrings to improve docstring coverage.
✅ Passed checks (2 passed)
Check name Status Explanation
Title check ✅ Passed The title clearly and concisely summarizes the main change: passing a bash script to srun instead of the command directly, which is the primary modification described in the changeset.
Description Check ✅ Passed Check skipped - CodeRabbit’s high-level summary is enabled.
✨ Finishing touches
  • 📝 Generate docstrings
🧪 Generate unit tests (beta)
  • Create PR with unit tests
  • Post copyable unit tests in a comment

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

Comment @coderabbitai help to get the list of available commands and usage tips.

Copy link

@coderabbitai coderabbitai bot left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Actionable comments posted: 1

🧹 Nitpick comments (2)
snakemake_executor_plugin_slurm_jobstep/__init__.py (2)

129-130: Consider limiting script logging for very large scripts.

While logging the script content is valuable for debugging, very large scripts could bloat logs. Consider logging the script length or first few lines instead.

Apply this diff to log script metadata instead:

-        srun_script_str = '\n' + srun_script if srun_script else 'N/A'
-        self.logger.debug(f"The script for this job is: {srun_script_str}")
+        if srun_script:
+            script_lines = srun_script.count('\n') + 1
+            script_preview = srun_script[:200] + ('...' if len(srun_script) > 200 else '')
+            self.logger.debug(f"The script for this job ({script_lines} lines, {len(srun_script)} chars): {script_preview}")
+        else:
+            self.logger.debug("The script for this job is: N/A")

133-133: Static analysis warning is a false positive, but consider conditional stdin.

The static analysis warning about shell=True (S602) is a false positive here. This change actually improves security by moving the job script (user-controlled content) from the shell command line to stdin. The call variable is constructed entirely from controlled sources.

However, stdin=subprocess.PIPE is set for all jobs but only used for SMP jobs. Consider setting it conditionally for minor efficiency improvement.

Apply this diff to conditionally set stdin:

-        jobsteps[job] = subprocess.Popen(call, shell=True, text=True, stdin=subprocess.PIPE)
+        stdin_arg = subprocess.PIPE if srun_script is not None else None
+        jobsteps[job] = subprocess.Popen(
+            call, shell=True, text=True, stdin=stdin_arg
+        )
📜 Review details

Configuration used: CodeRabbit UI

Review profile: CHILL

Plan: Pro

📥 Commits

Reviewing files that changed from the base of the PR and between 901c000 and 2f9d19b.

📒 Files selected for processing (1)
  • snakemake_executor_plugin_slurm_jobstep/__init__.py (2 hunks)
🧰 Additional context used
📓 Path-based instructions (1)
**/*.py

⚙️ CodeRabbit configuration file

**/*.py: Do not try to improve formatting.
Do not suggest type annotations for functions that are defined inside of functions or methods.
Do not suggest type annotation of the self argument of methods.
Do not suggest type annotation of the cls argument of classmethods.
Do not suggest return type annotation if a function or method does not contain a return statement.

Files:

  • snakemake_executor_plugin_slurm_jobstep/__init__.py
🪛 Ruff (0.14.6)
snakemake_executor_plugin_slurm_jobstep/__init__.py

133-133: subprocess call with shell=True identified, security issue

(S602)

🔇 Additional comments (2)
snakemake_executor_plugin_slurm_jobstep/__init__.py (2)

61-61: LGTM!

The introduction of srun_script to hold the bash script is clear and follows the PR objective of passing scripts via stdin.


122-124: Clarification needed: SLURM stdin forwarding limitation may affect this approach.

The web search revealed that "Slurm job scripts do not get stdin forwarded — if your rule expects stdin (bash reading from stdin), write input to a file (or use here-doc / bash -c "…") inside the job script". This directly challenges the assumption that bash -s will read the script from stdin in a SLURM context.

Before approving this change, confirm:

  1. How is srun_script actually passed to bash via stdin if SLURM doesn't forward stdin by default?
  2. Is there explicit configuration or a workaround being used that enables stdin forwarding?
  3. Verify this matches the approach in the referenced upstream PR #380

Without clarification on how stdin is actually provided to bash within the SLURM job execution context, the logic of storing the script in srun_script and using bash -s may not work as intended.

Comment on lines +134 to +137
if srun_script is not None:
# pass the srun bash script via stdin
jobsteps[job].stdin.write(srun_script)
jobsteps[job].stdin.close()
Copy link

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

⚠️ Potential issue | 🟠 Major

Add error handling for stdin write operation.

Writing to stdin can raise BrokenPipeError if the subprocess terminates prematurely. Without error handling, this would cause an unhandled exception with unclear error messaging. Wrap the write operation in try-except for better error reporting and robustness.

Apply this diff to add error handling:

         if srun_script is not None:
-            # pass the srun bash script via stdin
-            jobsteps[job].stdin.write(srun_script)
-            jobsteps[job].stdin.close()
+            try:
+                # pass the srun bash script via stdin
+                jobsteps[job].stdin.write(srun_script)
+                jobsteps[job].stdin.close()
+            except BrokenPipeError:
+                # subprocess terminated before reading stdin
+                self.logger.error(
+                    f"Failed to write script to stdin for job {job}. "
+                    "Subprocess may have terminated prematurely."
+                )
+                self.report_job_error(SubmittedJobInfo(job))
+                raise WorkflowError(
+                    f"Job {job} failed: subprocess terminated before reading script"
+                )
📝 Committable suggestion

‼️ IMPORTANT
Carefully review the code before committing. Ensure that it accurately replaces the highlighted code, contains no missing lines, and has no issues with indentation. Thoroughly test & benchmark the code to ensure it meets the requirements.

Suggested change
if srun_script is not None:
# pass the srun bash script via stdin
jobsteps[job].stdin.write(srun_script)
jobsteps[job].stdin.close()
if srun_script is not None:
try:
# pass the srun bash script via stdin
jobsteps[job].stdin.write(srun_script)
jobsteps[job].stdin.close()
except BrokenPipeError:
# subprocess terminated before reading stdin
self.logger.error(
f"Failed to write script to stdin for job {job}. "
"Subprocess may have terminated prematurely."
)
self.report_job_error(SubmittedJobInfo(job))
raise WorkflowError(
f"Job {job} failed: subprocess terminated before reading script"
)
🤖 Prompt for AI Agents
In snakemake_executor_plugin_slurm_jobstep/__init__.py around lines 134-137, the
direct write to jobsteps[job].stdin can raise BrokenPipeError if the subprocess
exits; wrap the write and close in a try/except/finally block that catches
BrokenPipeError (and optionally OSError), logs a clear error message including
the exception details using the module's existing logger (e.g., logger or
process_logger), and ensures stdin is closed in finally to avoid resource leaks;
do not swallow unexpected exceptions unless re-raised after logging.

@dmwnz dmwnz changed the title feat: pass a bash script to srun rather than the command directly to circumvent SLURM command-line character limit feat: pass a bash script to srun rather than the command directly Nov 27, 2025
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant