Use `/usr/bin/env -S bash` instead of `/bin/bash` on the host system #5684

rollf · 2025-01-20T12:51:06Z

On systems without /bin/bash (e.g. NixOS), invoking tasks may fail because of some hard-coded references to /bin/bash. Those references are either

fine (because the task runs in a container where /bin/bash is available),
don't matter (because the shebang is ignored)
or cause failure (if /bin/bash cannot be invoked).

This PR adapts Nextflow for the last scenario. Instead of invoking /bin/bash, /usr/bin/env -S bash is used where necessary. This approach (or similar) has been discussed previously:

However, compared the previous discussions/implementations, I specifically do not propose to change /bin/bash unless really necessary. By "necessary" I mean that bash is invoked on a local host system where it may not be available. Namely, when <task>/.command.run shall be invoked (always on the host) and when <task>/.command.sh is invoked in a non-container context (i.e. also on the host). I have not touched any other occurrences of /bin/bash, esp. not within the container.

I'm happy to change the the code as requested. I have not added any comments since I expect (at least) change requests. Please review this with an unbiased mind set. Multiple people have expressed their intention to use nextflow on systems without /bin/bash.

I have tested solely on NixOS.

netlify · 2025-01-20T12:51:28Z

✅ Deploy Preview for nextflow-docs-staging canceled.

Name	Link
🔨 Latest commit	`3b55cc6`
🔍 Latest deploy log	https://app.netlify.com/sites/nextflow-docs-staging/deploys/67907cf0c510180008a52d85

pditommaso · 2025-01-20T15:39:14Z

modules/nextflow/src/main/groovy/nextflow/executor/BashWrapperBuilder.groovy

@@ -85,6 +87,7 @@ class BashWrapperBuilder {
            log.warn "Invalid value for `NXF_DEBUG` variable: $str -- See http://www.nextflow.io/docs/latest/config.html#environment-variables"
        }
        BASH = Collections.unmodifiableList( level > 0 ? ['/bin/bash','-uex'] : ['/bin/bash','-ue'] )
+        ENV_BASH = Collections.unmodifiableList(['/usr/bin/env', '-S', 'bash', BASH[1]])


Why introducing this variable instead of relying on process.shell?

Thanks for looking into this. process.shell is /bin/bash ... by default. Does this answer your question?

My point is that after #4292 you should be able to use process.shell = ['/usr/bin/env, '-S', 'bash'] in your config.

Sorry, I still need clarification: Do you suggest this PR is not necessary at all because one could set process.shell = ['/usr/bin/env, '-S', 'bash'] in a pipeline's configuration or do you suggest a different implementation (not using ENV_BASH as I have proposed) would be a technically better approach to achieve this PR's goal?

"Yes", to my first or second interpretation? I feel like the are mutually exclusive. 🤷

Yes, I suggest this PR is not necessary at all because one could set process.shell

[Disclaimer: I assume that the additional usage of /usr/bin/env -S bash on the host system (to invoke .command.run as well as for non-container task execution) is acceptable by the Nextflow maintainers and we just need to sort out the technical implementation. Note that the POSIX-compliant existence of /usr/bin/env is already expected by Nextflow and please also consider the PR description above.]

procoess.shell cannot not be used for various reasons:

The value of process.shell will be used for task execution in both the non-container as well as the container scenario. This means that one would need to set this to a value that works in the host as well as in the container env if one wants to freely disable/enable containerization. And this contradicts the whole point.

Setting process.shell =/usr/bin/env -S bash -ue (because I want -ue) in the nextflow.config causes problems during tracing (because due to Use shell directive with trace command in wrapper script #4292, Nextflow will favor process.shell's actual value [it's not BASH] and then use -ue which Nextflow specifically doesn't expect for tracing)

Even if it worked, that means that everybody without /bin/bash needs to adapt every pipeline they want to run. I'm not trying to fix a specific problem I have with a specific pipeline, I'd like to contribute here in order to solve the general problem.

The use of /usr/bin/env is not viable because adds another level of indirection of in the process execution making (even) more complex the composition of the launch command and the passing of shell option.

I propose to address the problem of providing a custom Bash path in a more structural manner in this PR #5696

rollf · 2025-01-20T12:54:49Z

modules/nextflow/src/main/groovy/nextflow/executor/BashWrapperBuilder.groovy

@@ -65,6 +65,8 @@ class BashWrapperBuilder {

    static final public List<String> BASH


I suggest to rename BASH to BIN_BASH to make the difference clear.

rollf · 2025-01-20T12:55:32Z

modules/nextflow/src/main/groovy/nextflow/executor/local/LocalTaskHandler.groovy

@@ -131,7 +131,7 @@ class LocalTaskHandler extends TaskHandler implements FusionAwareTask {
    }

    protected ProcessBuilder localProcessBuilder() {
-        final cmd = new ArrayList<String>(BashWrapperBuilder.BASH) << wrapperFile.getName()
+        final cmd = new ArrayList<String>(BashWrapperBuilder.ENV_BASH) << wrapperFile.getName()


This makes sure that .command.run gets executed correctly on a host without /bin/bash.

rollf · 2025-01-20T12:56:27Z

modules/nextflow/src/main/groovy/nextflow/processor/TaskBean.groovy

@@ -136,7 +136,7 @@ class TaskBean implements Serializable, Cloneable {
        this.useMicromamba = task.getCondaConfig()?.useMicromamba()
        this.spackEnv = task.getSpackEnv()
        this.moduleNames = task.config.getModule()
-        this.shell = task.config.getShell() ?: BashWrapperBuilder.BASH
+        this.shell = task.config.getShell(task.isContainerEnabled()) ?: BashWrapperBuilder.BASH


At this point, task will always (?) have a shell since it comes with the process defaults. I have not touched the "else"-block here since I think it is actually not necessary (and it did not cause issues for me even though I don't have /bin/bash).

rollf · 2025-01-20T12:56:40Z

modules/nextflow/src/main/groovy/nextflow/processor/TaskConfig.groovy

@@ -374,10 +374,10 @@ class TaskConfig extends LazyMap implements Cloneable {
        return (List<String>) get('secret')
    }

-    List<String> getShell() {
+    List<String> getShell(boolean isContainerEnabled = true) {


I'm happy to adapt the unit tests but I don't know how 🤷 .

rollf · 2025-01-21T07:32:34Z

modules/nextflow/src/main/groovy/nextflow/executor/BashWrapperBuilder.groovy

@@ -85,6 +87,7 @@ class BashWrapperBuilder {
            log.warn "Invalid value for `NXF_DEBUG` variable: $str -- See http://www.nextflow.io/docs/latest/config.html#environment-variables"
        }
        BASH = Collections.unmodifiableList( level > 0 ? ['/bin/bash','-uex'] : ['/bin/bash','-ue'] )
+        ENV_BASH = Collections.unmodifiableList(['/usr/bin/env', '-S', 'bash', BASH[1]])


Thanks for looking into this. process.shell is /bin/bash ... by default. Does this answer your question?

On systems without /bin/bash (e.g. NixOS), invoking tasks may fail because of some hard-coded references to /bin/bash. Those references are either fine (because the task runs in a container where /bin/bash *is* available), don't matter (because the shebang is ignored) or cause failure (because /bin/bash cannot be invoked). This commit adapts Nextflow for the last scenario. Instead of invoking `/bin/bash`, the actually available bash executable is called. Signed-off-by: Rolf Schröder <rolf.schroeder@limbus-medtec.com>

rollf · 2025-01-22T05:09:44Z

modules/nextflow/src/main/groovy/nextflow/script/ProcessConfig.groovy

@@ -125,7 +125,6 @@ class ProcessConfig implements Map<String,Object>, Cloneable {
    static final Map<String,Object> DEFAULT_CONFIG = [
            debug: false,
            cacheable: true,
-            shell: BashWrapperBuilder.BASH,


The changes in this file and the next one make sure that Nextflow does not assume /bin/bash execution by default. Rather, it used /bin/bash when the configuration requests task containerization, otherwise, /usr/bin/env -S bash is used (i.e. when the task is to be executed on the host system itself).

pditommaso · 2025-01-22T13:51:50Z

Closing in favour of #5696

pditommaso force-pushed the master branch 2 times, most recently from 1f834a0 to 6454605 Compare January 20, 2025 15:13

pditommaso reviewed Jan 20, 2025

View reviewed changes

rollf commented Jan 21, 2025

View reviewed changes

rollf changed the title ~~Allow Nextflow to be used on (host) systems without /bin/bash~~ Use /usr/bin/env -S bash instead of /bin/bash on the host system Jan 21, 2025

rollf marked this pull request as ready for review January 21, 2025 12:38

rollf mentioned this pull request Jan 21, 2025

Make Nextflow compatible with NixOS (or other systems without /bin/bash) #5432

Closed

bentsherman added the software/bash label Jan 21, 2025

rollf force-pushed the nextflow-env-bash branch from a1eb9a5 to 3b55cc6 Compare January 22, 2025 05:06

rollf commented Jan 22, 2025

View reviewed changes

This was referenced Jan 22, 2025

Add newlines and set command to fix #3416 nf-core/tools#3417

Open

Template: Fix process.shell in nextflow.config nf-core/tools#3416

Merged

pditommaso closed this Jan 22, 2025

rollf mentioned this pull request Jan 22, 2025

Make Bash path configurable #5696

Open

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Use `/usr/bin/env -S bash` instead of `/bin/bash` on the host system #5684

Use `/usr/bin/env -S bash` instead of `/bin/bash` on the host system #5684

rollf commented Jan 20, 2025 •

edited

Loading

netlify bot commented Jan 20, 2025 •

edited

Loading

pditommaso Jan 20, 2025

rollf Jan 21, 2025

pditommaso Jan 21, 2025

rollf Jan 21, 2025

pditommaso Jan 21, 2025

rollf Jan 21, 2025

pditommaso Jan 21, 2025

rollf Jan 21, 2025 •

edited

Loading

pditommaso Jan 22, 2025

rollf Jan 20, 2025

rollf Jan 20, 2025

rollf Jan 20, 2025

rollf Jan 20, 2025

rollf Jan 21, 2025

rollf Jan 22, 2025

pditommaso commented Jan 22, 2025

		@@ -65,6 +65,8 @@ class BashWrapperBuilder {

		static final public List<String> BASH

Use /usr/bin/env -S bash instead of /bin/bash on the host system #5684

Use /usr/bin/env -S bash instead of /bin/bash on the host system #5684

Conversation

rollf commented Jan 20, 2025 • edited Loading

netlify bot commented Jan 20, 2025 • edited Loading

✅ Deploy Preview for nextflow-docs-staging canceled.

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

rollf Jan 21, 2025 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

pditommaso commented Jan 22, 2025

Use `/usr/bin/env -S bash` instead of `/bin/bash` on the host system #5684

Use `/usr/bin/env -S bash` instead of `/bin/bash` on the host system #5684

rollf commented Jan 20, 2025 •

edited

Loading

netlify bot commented Jan 20, 2025 •

edited

Loading

rollf Jan 21, 2025 •

edited

Loading