use log class in list_down_nodes

scripts/pyfe/pyfe/list_down_nodes.py

@@ -5,7 +5,7 @@
 from time import time
 from pyfe import scr_const, scr_common
 from pyfe.list_dir import list_dir
-from pyfe.scr_common import runproc, pipeproc, scr_prefix
+from pyfe.scr_common import runproc, pipeproc
 
 # mark any nodes specified on the command line
 def remove_argument_excluded_nodes(nodes=[],nodeset_down=[]):
@@ -18,7 +18,7 @@ def remove_argument_excluded_nodes(nodes=[],nodeset_down=[]):
 
 # The main scr_list_down_nodes method.
 # this method takes an scr_env, the contained resource manager will determine which methods above to use
-def list_down_nodes(reason=False, free=False, nodeset_down='', log_nodes=False, runtime_secs=None, nodeset=None, scr_env=None):
+def list_down_nodes(reason=False, free=False, nodeset_down='', runtime_secs=None, nodeset=None, scr_env=None, log=None):
   if scr_env is None or scr_env.resmgr is None or scr_env.param is None:
     return 1
   bindir = scr_const.X_BINDIR
@@ -40,14 +40,6 @@ def list_down_nodes(reason=False, free=False, nodeset_down='', log_nodes=False,
   # get list of nodes from nodeset
   nodes = scr_env.resmgr.expand_hosts(nodeset)
 
-  # get prefix directory
-  prefix = scr_env.get_prefix()
-
-  # get jobid
-  jobid = resourcemgr.getjobid()
-  #if jobid == 'defjobid': # job id could not be determined
-  #  print('Could not determine the job id') # the only place this is used here is in the logging below
-
   ### In each of the scr_list_down_nodes.in
   ### these nodes are marked as unavailable, and also removed from the list to log
   ### There is no use to keep track of them in the unavailable dictionary
@@ -66,23 +58,23 @@ def list_down_nodes(reason=False, free=False, nodeset_down='', log_nodes=False,
   # TODO: read exclude list from a file, as well?
 
   # print any failed nodes to stdout and exit with non-zero
-  if len(unavailable)>0:
+  if len(unavailable) > 0:
     # log each newly failed node, along with the reason
-    if log_nodes:
-      # scr_common.log calls the external program: scr_log_event
-      # the method will also accept a dictionary (instead of a string)
-      # for the event_note argument, this moves the loop closer to the runproc call
-      scr_common.log(bindir=bindir, prefix=prefix, jobid=jobid, event_type='NODE_FAIL', event_note=unavailable, event_start=start_time, event_secs=runtime_secs)
+    if log:
+      for node in unavailable:
+        note = node + ": " + unavailable[node]
+        log.event('NODE_FAIL', note=note, start=start_time, secs=runtime_secs)
+
     # now output info to the user
-    ret=''
     if reason:
       # list each node and the reason each is down
+      reasons = []
       for node in unavailable:
-        ret += node+': '+unavailable[node]+'\n'
-      ret = ret[:-1] ### take off the final trailing newline (?)
+        reasons.append(node + ': ' + unavailable[node])
+      return "\n".join(reasons)
     else:
       # simply print the list of down node in range syntax
-      ret = scr_env.resmgr.compress_hosts(list(unavailable))
-    return ret
+      return scr_env.resmgr.compress_hosts(list(unavailable))
+
   # otherwise, don't print anything and exit with 0
   return 0

scripts/pyfe/pyfe/scr_list_down_nodes.py

@@ -14,6 +14,7 @@
 from pyfe.scr_environment import SCR_Env
 from pyfe.scr_param import SCR_Param
 from pyfe.resmgr import AutoResourceManager
+from pyfe.cli import SCRLog
 
 if __name__=='__main__':
   parser = argparse.ArgumentParser(add_help=False, argument_default=argparse.SUPPRESS, prog='scr_list_down_nodes')
@@ -31,5 +32,14 @@
     scr_env = SCR_Env()
     scr_env.resmgr = AutoResourceManager()
     scr_env.param = SCR_Param()
-    ret = list_down_nodes(reason=args['reason'], free=args['free'], nodeset_down=args['down'], log_nodes=args['log'], runtime_secs=args['secs'], nodeset=args['[nodeset]'], scr_env=scr_env)
+
+    # create log object if asked to log down nodes
+    log = None
+    if args['log']:
+      prefix = scr_env.get_prefix()
+      jobid = scr_env.resmgr.getjobid()
+      user = scr_env.get_user()
+      log = SCRLog(prefix, jobid, user=user)
+
+    ret = list_down_nodes(reason=args['reason'], free=args['free'], nodeset_down=args['down'], runtime_secs=args['secs'], nodeset=args['[nodeset]'], scr_env=scr_env, log=log)
     print(str(ret),end='') ### remove trailing newlines?

scripts/pyfe/pyfe/scr_run.py

@@ -210,7 +210,7 @@ def scr_run(launcher='',launcher_args=[],run_cmd='',restart_cmd='',restart_args=
       down_nodes = ''
     if down_nodes!='':
       # print the reason for the down nodes, and log them
-      list_down_nodes(reason=True, free=free_flag, nodeset_down=down_nodes, log_nodes=True, runtime_secs='0', scr_env=scr_env)
+      list_down_nodes(reason=True, free=free_flag, nodeset_down=down_nodes, runtime_secs='0', scr_env=scr_env, log=log)
 
       # if this is the first run, we hit down nodes right off the bat, make a record of them
       if attempts==0:
@@ -290,7 +290,7 @@ def scr_run(launcher='',launcher_args=[],run_cmd='',restart_cmd='',restart_args=
     run_secs = end_secs - start_secs
 
     # check for and log any down nodes
-    list_down_nodes(reason=True, nodeset_down=keep_down, log_nodes=True, runtime_secs=str(run_secs), scr_env=scr_env)
+    list_down_nodes(reason=True, nodeset_down=keep_down, runtime_secs=str(run_secs), scr_env=scr_env, log=log)
 
     # log stats on the latest run attempt
     log.event('RUN_END', note='run=' + str(attempts), start=str(end_secs), secs=str(run_secs))