Revert "[pytorch] Moving torch.compile worker process logs to a dedicated rank based log directory (#159874)"

This reverts commit 9fd5b5f73589cf08dca60910368cc0f05c7906c8.

Reverted https://github.com/pytorch/pytorch/pull/159874 on behalf of https://github.com/malfet due to Broke lint ([comment](https://github.com/pytorch/pytorch/pull/159874#issuecomment-3161896978))

test/inductor/test_compile_worker.py

@@ -1,8 +1,6 @@
 # Owner(s): ["module: inductor"]
-import importlib
 import operator
 import os
-import tempfile
 
 from torch._inductor.compile_worker.subproc_pool import (
     raise_testexc,
@@ -13,6 +11,7 @@ from torch._inductor.test_case import TestCase
 from torch.testing._internal.common_utils import skipIfWindows
 from torch.testing._internal.inductor_utils import HAS_CPU
 
+
 class TestCompileWorker(TestCase):
     @skipIfWindows(msg="pass_fds not supported on Windows.")
     def test_basic_jobs(self):
@@ -67,18 +66,6 @@ class TestCompileWorker(TestCase):
         finally:
             pool.shutdown()
 
-    @skipIfWindows(msg="pass_fds not supported on Windows.")
-    def test_logging(self):
-        os.environ["MAST_HPC_JOB_NAME"] = "test_job"
-        os.environ["ROLE_RANK"] = "0"
-        with tempfile.NamedTemporaryFile(delete=True) as temp_log:
-            os.environ["TORCHINDUCTOR_WORKER_LOGPATH"] = temp_log.name
-            pool = SubprocPool(2)
-            try:
-                pool.submit(operator.add, 100, 1)
-                self.assertEqual(os.path.exists(temp_log.name), True)
-            finally:
-                pool.shutdown()
 
 if __name__ == "__main__":
     from torch._inductor.test_case import run_tests

torch/_inductor/compile_worker/subproc_pool.py

@@ -145,24 +145,11 @@ class SubprocPool:
             f"--write-fd={str(subproc_write_fd)}",
             f"--torch-key={torch_key_str}",
         ]
-        mast_job_id = os.environ.get("MAST_HPC_JOB_NAME", None)
-        global_rank = os.environ.get("ROLE_RANK", "0")
-        worker_log_path = os.environ.get("TORCHINDUCTOR_WORKER_LOGPATH", config.worker_log_path)
-        stdout_pipe = None
-        stderr_pipe = None
-        self.log_file = None
-
-        if mast_job_id is not None:
-            log_loc = f"{worker_log_path}{global_rank}"
-            self.log_file = open(log_loc, "w")
-        elif config.worker_suppress_logging:
+        local = False
+        if config.worker_suppress_logging:
             log.info("Suppressing compile worker output due to config")
-            self.log_file = open(os.devnull, "w")
+            local = True
 
-        if self.log_file:
-            stdout_pipe = self.log_file
-            stderr_pipe = self.log_file
-        
         self.process = subprocess.Popen(
             cmd,
             env={
@@ -177,10 +164,9 @@ class SubprocPool:
                 "LD_LIBRARY_PATH": get_ld_library_path(),
             },
             pass_fds=(subproc_read_fd, subproc_write_fd),
-            stdout=stdout_pipe,
-            stderr=stderr_pipe,
+            stdout=subprocess.DEVNULL if local else None,
+            stderr=subprocess.DEVNULL if local else None,
         )
-
         self.write_lock = threading.Lock()
         self.read_thread = threading.Thread(
             target=self._read_thread, name="InductorSubproc", daemon=True
@@ -276,8 +262,6 @@ class SubprocPool:
                 _send_msg(self.write_pipe, MsgHeader.SHUTDOWN)
                 self.write_pipe.close()
             self.process.wait(300)
-            if self.log_file:
-                self.log_file.close()
         except OSError as e:
             log.warning("Ignored OSError in pool shutdown:  %s", e)
         finally:

torch/_inductor/config.py

@@ -81,9 +81,6 @@ disable_progress = True
 # Whether to enable printing the source code for each future
 verbose_progress = False
 
-# Configurable compile worker logging path for subproc_pool
-worker_log_path = "/logs/dedicated_log_torch_compile_worker_rank" if is_fbcode() else None
-
 # precompilation timeout
 precompilation_timeout_seconds: int = 60 * 60