Print stderr when both qsub and sacctmgr fails

The command `sacctmgr` fails on some computers (mila01 namely), but the
current behavior gives the impression sbatch is simply not available.

Printing the stderr makes it more obvious that sbatch should be
available, but something is broken behind sacctmgr. It only appears when
using -vv options nevertheless.

Author: bouthilx

smartdispatch/utils.py

@@ -2,12 +2,16 @@
 import distutils.spawn
 import hashlib
 import json
+import logging
 import re
 import unicodedata
 
 from subprocess import Popen, PIPE
 
 
+logger = logging.getLogger(__name__)
+
+
 TIME_REGEX = re.compile(
     "^(?:(?:(?:(\d*):)?(\d*):)?(\d*):)?(\d*)$")
 
@@ -157,11 +161,18 @@ def detect_cluster():
 
 def get_slurm_cluster_name():
     try:
-        stdout = Popen("sacctmgr list cluster", stdout=PIPE, shell=True).communicate()[0]
+        popen = Popen("sacctmgr list cluster", stdout=PIPE, shell=True)
+        stdout, stderr = popen.communicate()
+    except OSError:
+        return None
+
+    try:
         stdout = stdout.decode()
         cluster_name = stdout.splitlines()[2].strip().split(' ')[0]
-    except IndexError, OSError:
+    except IndexError:
+        logger.debug(stderr)
         return None
+
     return cluster_name