oracle db recovery: some fixes

ProductionTools/oracle_db_disaster_recovery/oracle_check_closed_ondisk.py

@@ -32,7 +32,7 @@ def oracle_check_closed_ondisk(oradb_url, run_number, outliers_file,
 
     ora_state_3 = oradb.db.curs.execute("""select SFOID, SFOPFN from SFO_TZ_FILE
             where RUNNR=:runnr and FILESTATE='CLOSED' and TRANSFERSTATE='ONDISK'
-            """, runnr=run_number).fetchall()
+            and SFOID like 'SFO%'""", runnr=run_number).fetchall()
     if verbose: print(f'Found {len(ora_state_3)} (closed,ondisk)')
 
     with open(outliers_file, 'w') as outf:

ProductionTools/oracle_db_disaster_recovery/oracle_check_closed_transferred.py

@@ -33,7 +33,7 @@ def oracle_check_closed_transferred(oradb_url, run_number, outliers_file,
     if report: reportf = open(report, 'a')
 
     ora_state_4 = oradb.db.curs.execute("""select SFOID, SFOPFN from SFO_TZ_FILE
-            where RUNNR=:runnr and FILESTATE='CLOSED'
+            where RUNNR=:runnr and FILESTATE='CLOSED' and SFOID like 'SFO%'
             and TRANSFERSTATE='TRANSFERRED'""", runnr=run_number).fetchall()
     if verbose: print(f'Found {len(ora_state_4)} (closed,transferred)')
 

ProductionTools/oracle_db_disaster_recovery/oracle_db_disaster_recovery.py

@@ -114,7 +114,7 @@ for i in args.sfos:
     if not os.path.exists(f'sfo-{i}.{args.run_number}.sqlite'):
         scp_cmd = ['scp', f'pc-tdq-sfo-{i}:/dsk1/sqlite/sfo-sqlite-{args.run_number}.db',
                 f'sfo-{i}.{args.run_number}.sqlite']
-        subprocess.run(scp_cmd, check=True)
+        subprocess.run(scp_cmd, check=True, stdout=subprocess.PIPE)
         if args.verbose: print(f'downloaded sqlite: sfo-{i}.{args.run_number}.sqlite')
 
 if args.verbose: print('recovery step A: bring state 1 (missing) and 2 (opened) to 3 (closed)')
@@ -134,6 +134,10 @@ oracle_check_closed_ondisk(args.oradb_url, args.run_number, args.outliers_file,
         verbose=args.verbose, dryrun=args.dryrun, report=args.report)
 
 if os.path.exists(args.outliers_file) and os.path.getsize(args.outliers_file) != 0:
+    print(f'there are some outliers: check file "{args.outliers_file}"')
+    print('CastorScript instances NOT restarted: re-enable them after having'
+            ' taken care of outliers')
+else:
     if args.verbose: print('no outliers')
     if args.verbose: print('starting CastorScript instances')
     if not args.dryrun:
@@ -141,7 +145,3 @@ if os.path.exists(args.outliers_file) and os.path.getsize(args.outliers_file) !=
             os.rename(f'/mnt/daq_area_rw/castor/pc-tdq-sfo-{i}/atlascdr/prod.stopped',
                     f'/mnt/daq_area_rw/castor/pc-tdq-sfo-{i}/atlascdr/prod.cfg')
             if args.verbose: print(f're-enabled CS on sfo-{i}')
-else:
-    print(f'there are some outliers: check file "{args.outliers_file}"')
-    print('CastorScript instances NOT restarted: re-enable them after having'
-            ' taken care of outliers')

Script/cs/Tools/MigrationCheckers/MergedMigrated.py

-#!/bin/env python
-import os.path
-import sys
-
-if __name__ == "__main__":
-    from os.path import dirname, abspath, join
-
-    # add /../../CastorScript/Script to path so imports keeps working
-    SCRIPT_DIR = abspath(join(dirname(__file__), '..','..',".."))
-    sys.path.append(SCRIPT_DIR)
-
-from cs.Tools.FilenameParsers.SFOFileNameParser import SFOFileNameParser
-
-BASEDIR = '/castor/cern.ch/grid/atlas/tzero/prod1/perm/'
-
-
-def MergedMigrated(castorfile, stager, pool, backend, logger=None, verbose=False):
-    """
-Destination directory:
-/<basedir>/<projecttag>/<streamtype_streamname>/<runnr-7digit>/<dataset>/
-
-basedir = /castor/cern.ch/grid/atlas/tzero/prod1/perm/
-projecttag = data09_cos
-dataset = data09_cos.00122050.physics_IDCosmic.merge.RAW
-
-Merged file name is:
-
-1) if all SFO files of a LB fit into a single file
-  <project>.<runnr>.<stream>.daq.RAW._lbXXXX._SFO-N._<fileseqnr>.data, N=4,5,9,10
-  --> <project>.<runnr>.<stream>.merge.RAW._lbXXXX._SFO-ALL._0001.<attemptnr>
-
-2) if all SFO files of a LB don't fit into a single file
-  <project>.<runnr>.<stream>.daq.RAW._lbXXXX._SFO-N._<fileseqnr>.data, N=4,5,9,10
-  --> <project>.<runnr>.<stream>.merge.RAW._lbXXXX._SFO-N._<fileseqnr>.<attemptnr>, N=4,5,9,10
-  """
-    ### Build the target directory name
-    parsed = SFOFileNameParser(os.path.basename(castorfile))
-
-    dataset = '%s.%s.%s_%s.merge.RAW' \
-              % (parsed.ProjectTag(), parsed.RunNr(), \
-                 parsed.StreamType(), parsed.StreamName())
-
-
-    path = os.path.join(BASEDIR, parsed.ProjectTag(), \
-                        '%s_%s' % (parsed.StreamType(), parsed.StreamName()), \
-                        '%s' % parsed.RunNr(), \
-                        '%s' % dataset)
-    if verbose:
-        print(path)
-
-    ### Look for files into the dataset dir
-    success, all_files = backend.listdir(path, stager, pool, logger)
-
-    ### If fails, return false
-    if not success: return False
-
-    ### Create the two type of name we expect (neglecting attempt number)
-    notmerged = os.path.splitext(os.path.basename(castorfile))[0]
-    notmerged = notmerged.replace('.daq.','.merge.')
-    merged = '.'.join(notmerged.split('.',6)[:-1]+['_SFO-ALL._0001'])
-
-    ### Find all the files with the correct name
-    files = [f for f in all_files
-             if f and (notmerged in f or merged in f)]
-
-    if not files: return False
-
-    ### Take the file with the highest attempt number
-    filename = sorted([(int(f.rsplit('.',1)[1]),f) for f in files])[-1][1]
-
-    ### Check the migration status for the file
-    return backend.migrated(os.path.join(path, filename),
-                            stager, pool, logger)
-
-if __name__ == '__main__':
-
-    import importlib
-    backend_module = importlib.import_module("cs.StorageBackends.castorstorage")
-
-    print(MergedMigrated(sys.argv[1], sys.argv[2], sys.argv[3],
-                         backend_module, verbose=True))