reformat PreJob. Fix #8938 (#8949)

src/python/ServerUtilities.py

@@ -1168,3 +1168,22 @@ def getRucioAccountFromLFN(lfn):
             return "{0}_group".format(account)
         return account
     raise Exception("Expected /store/(user,group)/rucio/<account>, got {0}".format(lfn))
+
+
+def pythonListToClassAdExprTree(aList):
+    # arg: aList : list of strings
+    # return: value : a string representing the classAd value
+    # transforms a python list of strings into a string containing the
+    # internal (ExprTree) representation of a ClassAd such that the ad
+    # will be converted to a python list when used like : myList = ad['myAd']
+    # https://htcondor.readthedocs.io/en/latest/apis/python-bindings/api/version2/classad2/classad.html#classad2.ClassAd
+    # inside python we can simply do ad['myAd'] = myList but if we want
+    # to create the ad in a submitted job via the JDL MY.<attribute> command
+    # https://htcondor.readthedocs.io/en/latest/man-pages/condor_submit.html#submit-description-file-commands
+    # the python list [a, b, c] needs to become '{"a","b","c"}'
+    # i.e. a string with the ExprTree representatio of a ClassAd list
+    import json
+    quotedItems = json.dumps(aList)  # from [s1, s2] to the string '["s1","s2"]'
+    quotedItems = quotedItems.lstrip('[').rstrip(']')  # remove square brackets [ ]
+    value = "{" + quotedItems + "}"  # make final string adding the curly brackets { }
+    return value

src/python/TaskWorker/Actions/DagmanCreator.py

@@ -21,7 +21,7 @@
 from ast import literal_eval
 
 from ServerUtilities import MAX_DISK_SPACE, MAX_IDLE_JOBS, MAX_POST_JOBS, TASKLIFETIME
-from ServerUtilities import getLock, checkS3Object
+from ServerUtilities import getLock, checkS3Object, pythonListToClassAdExprTree
 
 import TaskWorker.DataObjects.Result
 from TaskWorker.Actions.TaskAction import TaskAction
@@ -340,17 +340,16 @@ def makeJobSubmit(self, task):
         jobSubmit['My.CRAB_PublishDBSURL'] = classad.quote(task['tm_publish_dbs_url'])
         jobSubmit['My.CRAB_ISB'] = classad.quote(task['tm_cache_url'])
 
-        def pythonListToClassAdValue(aList):
-            # python lists need special handling to become the string '{"a","b",...,"c"}'
-            quotedItems = json.dumps(aList)  # from [s1, s2] to the string '["s1","s2"]'
-            quotedItems = quotedItems.lstrip('[').rstrip(']')  # remove square brackets [ ]
-            value = "{" + quotedItems + "}"  # make final string adding the curly brackets { }
-            return value
-        jobSubmit['My.CRAB_SiteBlacklist'] = pythonListToClassAdValue(task['tm_site_blacklist'])
-        jobSubmit['My.CRAB_SiteWhitelist'] =  pythonListToClassAdValue(task['tm_site_whitelist'])
-        jobSubmit['My.CRAB_AdditionalOutputFiles'] = pythonListToClassAdValue(task['tm_outfiles'])
-        jobSubmit['My.CRAB_EDMOutputFiles'] =  pythonListToClassAdValue(task['tm_edm_outfiles'])
-        jobSubmit['My.CRAB_TFileOutputFiles'] = pythonListToClassAdValue(task['tm_outfiles'])
+
+        # note about Lists
+        # in the JDL everything is a string, we can't use the simple classAd[name]=somelist
+        # but need the ExprTree format (what classAd.lookup() would return)
+        jobSubmit['My.CRAB_SiteBlacklist'] = pythonListToClassAdExprTree(task['tm_site_blacklist'])
+        jobSubmit['My.CRAB_SiteWhitelist'] =  pythonListToClassAdExprTree(task['tm_site_whitelist'])
+        jobSubmit['My.CRAB_AdditionalOutputFiles'] = pythonListToClassAdExprTree(task['tm_outfiles'])
+        jobSubmit['My.CRAB_EDMOutputFiles'] =  pythonListToClassAdExprTree(task['tm_edm_outfiles'])
+        jobSubmit['My.CRAB_TFileOutputFiles'] = pythonListToClassAdExprTree(task['tm_outfiles'])
+
         jobSubmit['My.CRAB_UserDN'] = classad.quote(task['tm_user_dn'])
         jobSubmit['My.CRAB_UserHN'] = classad.quote(task['tm_username'])
         jobSubmit['My.CRAB_AsyncDest'] = classad.quote(task['tm_asyncdest'])
@@ -709,7 +708,6 @@ def prepareJobArguments(self, dagSpecs, task):
             argDict['eventsPerLumi'] = task['tm_events_per_lumi']  #
             argDict['maxRuntime'] = dagspec['maxRuntime']  # -1
             argDict['scriptArgs'] = task['tm_scriptargs']
-            argDict['CRAB_AdditionalOutputFiles'] = "{}"
             # following one are for bkw compat. with CRABClient v3.241218 or earlier, to be removed
             argDict['CRAB_Archive'] = argDict['userSandbox']
             argDict['CRAB_ISB'] = 'dummy'

src/python/TaskWorker/Actions/DagmanResubmitter.py

@@ -7,7 +7,7 @@
 
 import HTCondorLocator
 
-from ServerUtilities import FEEDBACKMAIL
+from ServerUtilities import FEEDBACKMAIL, pythonListToClassAdExprTree
 from TaskWorker.Actions.TaskAction import TaskAction
 from TaskWorker.Actions.DagmanSubmitter import checkMemoryWalltime
 from TaskWorker.WorkerExceptions import TaskWorkerException
@@ -70,7 +70,6 @@ def executeInternal(self, *args, **kwargs): #pylint: disable=unused-argument
 
         ## Calculate new parameters for resubmitted jobs. These parameters will
         ## be (re)written in the _CONDOR_JOB_AD when we do schedd.edit() below.
-        ad = classad.ClassAd()
         params = {'CRAB_ResubmitList'  : 'jobids',
                   'CRAB_SiteBlacklist' : 'site_blacklist',
                   'CRAB_SiteWhitelist' : 'site_whitelist',
@@ -79,13 +78,7 @@ def executeInternal(self, *args, **kwargs): #pylint: disable=unused-argument
                   'CRAB_RequestedCpus'        : 'numcores',
                   'JobPrio'            : 'priority'
                  }
-        for taskparam in params.values():
-            if ('resubmit_'+taskparam in task) and task['resubmit_'+taskparam] is not None:
-                # the following seems to Stefano a very complicated way to handle
-                # in different (proper) way the cases where resubmit_xxxx is list and
-                # where it is a number. Should be verified and simplified
-                if isinstance(task['resubmit_'+taskparam], list):
-                    ad[taskparam] = task['resubmit_'+taskparam]
+
         if ('resubmit_jobids' in task) and task['resubmit_jobids']:
             self.logger.debug("Resubmitting when JOBIDs were specified")
             try:
@@ -94,12 +87,18 @@ def executeInternal(self, *args, **kwargs): #pylint: disable=unused-argument
                 # all the jobs, not only the ones we want to resubmit. That's why the pre-job
                 # is saving the values of the parameters for each job retry in text files (the
                 # files are in the directory resubmit_info in the schedd).
+                # We use classAds here as way to pass informations to PreJobs.
+                # Value in schedd.exit(const, ad, value) can be string or ExprTree
+                # https://htcondor.readthedocs.io/en/latest/apis/python-bindings/api/version2/htcondor2/schedd.html#htcondor2.Schedd.edit
+                # We need ExprTree when the ad correspond to a python list
                 for adparam, taskparam in params.items():
-                    if taskparam in ad:
-                        schedd.edit(rootConst, adparam, ad.lookup(taskparam))
-                    elif task['resubmit_' + taskparam] is not None:
-                        # param values set in this case are numbers, need to convert to strings
-                        schedd.edit(rootConst, adparam, str(task['resubmit_' + taskparam]))
+                    if task['resubmit_' + taskparam] is not None:
+                        if isinstance(task['resubmit_'+taskparam], list):
+                            newAdValue = pythonListToClassAdExprTree(task['resubmit_'+taskparam])
+                        else:
+                            newAdValue = str(task['resubmit_'+taskparam])
+                        schedd.edit(rootConst, adparam, newAdValue)
+                # finally restart the dagman with the 3 lines below
                 schedd.act(htcondor.JobAction.Hold, rootConst)
                 schedd.edit(rootConst, "HoldKillSig", 'SIGUSR1')
                 schedd.act(htcondor.JobAction.Release, rootConst)