Allow multiple docker experiments with same task names to run at the same time (#124)

hemildesai · web-flow · commit 283c0a13ec4a · 2024-12-16T11:52:36.000-08:00
Signed-off-by: Hemil Desai &lt;hemild@nvidia.com&gt;
diff --git a/src/nemo_run/run/torchx_backend/schedulers/docker.py b/src/nemo_run/run/torchx_backend/schedulers/docker.py
@@ -88,7 +88,10 @@ def _submit_dryrun(self, app: AppDef, cfg: Executor) -> AppDryRunInfo[DockerJobR
             cmd = [role.entrypoint] + role.args
             containers.append(
                 DockerContainer(
-                    name=role.name, command=cmd, executor=_current_executor, extra_env=role.env
+                    name=f"{executor.experiment_id}_{role.name}",
+                    command=cmd,
+                    executor=_current_executor,
+                    extra_env=role.env,
                 )
             )
 

Original file line number	Diff line number	Diff line change
`@@ -88,7 +88,10 @@ def _submit_dryrun(self, app: AppDef, cfg: Executor) -> AppDryRunInfo[DockerJobR`
`88`	`88`	`cmd = [role.entrypoint] + role.args`
`89`	`89`	`containers.append(`
`90`	`90`	`DockerContainer(`
`91`		`- name=role.name, command=cmd, executor=_current_executor, extra_env=role.env`
	`91`	`+ name=f"{executor.experiment_id}_{role.name}",`
	`92`	`+ command=cmd,`
	`93`	`+ executor=_current_executor,`
	`94`	`+ extra_env=role.env,`
`92`	`95`	`)`
`93`	`96`	`)`
`94`	`97`