rda-python-dsquasar 2.0.12__tar.gz → 3.0.3__tar.gz

rda_python_dsquasar-3.0.3/MANIFEST.in

@@ -0,0 +1 @@
+include src/rda_python_dsquasar/*.usg

rda_python_dsquasar-3.0.3/PKG-INFO

@@ -0,0 +1,87 @@
+Metadata-Version: 2.4
+Name: rda_python_dsquasar
+Version: 3.0.3
+Summary: RDA Python package to backup and recover RDA data archives to and from GLOBUS Quasar backup server
+Author-email: Zaihua Ji <zji@ucar.edu>
+Project-URL: Homepage, https://github.com/NCAR/rda-python-dsquasar
+Classifier: Programming Language :: Python :: 3
+Classifier: License :: OSI Approved :: MIT License
+Classifier: Operating System :: OS Independent
+Classifier: Development Status :: 5 - Production/Stable
+Requires-Python: >=3.7
+Description-Content-Type: text/markdown
+License-File: LICENSE
+Requires-Dist: rda_python_common
+Requires-Dist: rda_python_dsarch
+Dynamic: license-file
+
+# rda_python_dsquasar
+
+RDA Python package to back up and recover RDA data archives to and from the
+GLOBUS Quasar backup server.
+
+## Programs
+
+The package installs the following command-line utilities, all running as the
+current user:
+
+- `dsquasar` — back up and restore RDA dataset archives via Quasar
+- `tacctar` — create tar bundles for Quasar archival
+- `taccrec` — recover content from Quasar tar bundles
+
+Run any program with `--help` (or `-h`) for full usage details.
+
+## Environment setup
+
+Create a Python environment first; package installs in the next section run
+inside whichever environment you activate here.
+
+### Option A — Python venv (DECS machines)
+
+```bash
+python3 -m venv $ENVHOME          # e.g. /glade/u/home/gdexdata/gdexmsenv
+source $ENVHOME/bin/activate
+```
+
+### Option B — Conda (DAV/Casper)
+
+```bash
+conda create --prefix $ENVHOME python=3.12   # e.g. /glade/work/gdexdata/conda-envs/pg-gdex
+conda activate $ENVHOME
+```
+
+## Installing rda-python-dsquasar
+
+Pick whichever install mode fits your workflow.  All variants pull in the
+transitive dependencies (`rda_python_common`, `rda_python_dsarch`)
+automatically.
+
+For local development, clone this repo alongside your project and install it
+in editable mode so that changes are picked up without re-installing:
+
+```bash
+git clone https://github.com/NCAR/rda-python-dsquasar.git
+cd rda-python-dsquasar
+pip install -e .
+```
+
+To test a specific branch (e.g. an in-progress feature or fix branch), pass
+`-b/--branch` to `git clone`:
+
+```bash
+git clone -b <branch-name> https://github.com/NCAR/rda-python-dsquasar.git
+cd rda-python-dsquasar
+pip install -e .
+```
+
+For a regular (non-editable) install from a checkout:
+
+```bash
+pip install /path/to/rda-python-dsquasar
+```
+
+For a production install on a system that uses the published distribution:
+
+```bash
+pip install rda_python_dsquasar
+```

rda_python_dsquasar-3.0.3/README.md

@@ -0,0 +1,70 @@
+# rda_python_dsquasar
+
+RDA Python package to back up and recover RDA data archives to and from the
+GLOBUS Quasar backup server.
+
+## Programs
+
+The package installs the following command-line utilities, all running as the
+current user:
+
+- `dsquasar` — back up and restore RDA dataset archives via Quasar
+- `tacctar` — create tar bundles for Quasar archival
+- `taccrec` — recover content from Quasar tar bundles
+
+Run any program with `--help` (or `-h`) for full usage details.
+
+## Environment setup
+
+Create a Python environment first; package installs in the next section run
+inside whichever environment you activate here.
+
+### Option A — Python venv (DECS machines)
+
+```bash
+python3 -m venv $ENVHOME          # e.g. /glade/u/home/gdexdata/gdexmsenv
+source $ENVHOME/bin/activate
+```
+
+### Option B — Conda (DAV/Casper)
+
+```bash
+conda create --prefix $ENVHOME python=3.12   # e.g. /glade/work/gdexdata/conda-envs/pg-gdex
+conda activate $ENVHOME
+```
+
+## Installing rda-python-dsquasar
+
+Pick whichever install mode fits your workflow.  All variants pull in the
+transitive dependencies (`rda_python_common`, `rda_python_dsarch`)
+automatically.
+
+For local development, clone this repo alongside your project and install it
+in editable mode so that changes are picked up without re-installing:
+
+```bash
+git clone https://github.com/NCAR/rda-python-dsquasar.git
+cd rda-python-dsquasar
+pip install -e .
+```
+
+To test a specific branch (e.g. an in-progress feature or fix branch), pass
+`-b/--branch` to `git clone`:
+
+```bash
+git clone -b <branch-name> https://github.com/NCAR/rda-python-dsquasar.git
+cd rda-python-dsquasar
+pip install -e .
+```
+
+For a regular (non-editable) install from a checkout:
+
+```bash
+pip install /path/to/rda-python-dsquasar
+```
+
+For a production install on a system that uses the published distribution:
+
+```bash
+pip install rda_python_dsquasar
+```

{rda_python_dsquasar-2.0.12 → rda_python_dsquasar-3.0.3}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "rda_python_dsquasar"
-version = "2.0.12"
+version = "3.0.3"
 authors = [
   { name="Zaihua Ji", email="zji@ucar.edu" },
 ]

{rda_python_dsquasar-2.0.12 → rda_python_dsquasar-3.0.3}/src/rda_python_dsquasar/ds_quasar.py

@@ -91,7 +91,7 @@ TFCOUNT = 100          # if file count is greater, use MINSIZE for tar file
 SUBLMTS = 2000        # file count limit for a sub-group
 
 PGBACK = {
-   'workdir' : "{}/{}/quasar_backup".format(PgLOG.PGLOG['GDEXWORK'], PgLOG.PGLOG['GDEXUSER']),
+   'workdir' : "{}/{}/quasar_backup".format(PgLOG.PGLOG['GDEXWORK'], PgLOG.PGLOG['COMMONUSER']),
    'mproc' : 1,
    'action' : CTACTS,
    'chgdays' : 0,
@@ -1092,7 +1092,7 @@ def evaluate_file_stat(dsid, cate, pgrec):
    chome = PgLOG.PGLOG['DECSHOME'] if cate == 's' else PgLOG.PGLOG['DSDHOME']
 
    if pgrec['locflag'] == 'O':
-      bucket = 'gdex-decsdata' if cate == 's' else PgLOG.PGLOC['OBJCTBKT']
+      bucket = 'gdex-decsdata' if cate == 's' else PgLOG.PGLOG['OBJCTBKT']
       ofile = PgLOG.join_paths(dsid, fname)
       info = PgFile.check_object_file(ofile, bucket, 1, ERRACT)
       if not info: return False

{rda_python_dsquasar-2.0.12 → rda_python_dsquasar-3.0.3}/src/rda_python_dsquasar/dsquasar.py

@@ -12,7 +12,7 @@
 #             to gather Web and Saved bfiles, tar them into larger (> 2GB) bfiles, and
 #             back them up onto Quasar Backup Servers, at the Globus end points of
 #             NCAR GDEX Quasar and/or NCAR GDEX Quasar Drdata, for Backup and/or
-#             desaster recovery copies, respectively. The file tarring and backing up
+#             disaster recovery copies, respectively. The file tarring and backing up
 #             processes are completed in utility program dsarch.
 #
 #    Github : https://github.com/NCAR/rda-python-dsquasar.git
@@ -84,8 +84,12 @@ class DsQuasar(PgCMD, PgSplit):
       self.ONESIZE = 20*self.PGLOG['ONEGBS']  # 20GB, minimal file size to tar a single file 
       self.TFCOUNT = 100          # if file count is greater, use self.MINSIZE for tar file
       self.SUBLMTS = 2000         # file count limit for a sub-group
+      self.MPLIMIT = 500          # tar file count per batch process for multi-processing
+      self.MPMAX = 8              # maximum number of batch processes for multi-processing
+      self.MAXRUNTIME = 23*3600   # 23 hours; stop before the 24-hour PBS walltime
+      self.ONEHOUR = 3600         # seconds; time headroom needed to finish before walltime
       self.PGBACK = {
-         'workdir' : "{}/{}/quasar_backup".format(self.PGLOG['GDEXWORK'], self.PGLOG['GDEXUSER']),
+         'workdir' : "{}/{}/quasar_backup".format(self.PGLOG['GDEXWORK'], self.PGLOG['COMMONUSER']),
          'mproc' : 1,
          'action' : self.CTACTS,
          'chgdays' : 0,
@@ -97,6 +101,8 @@ class DsQuasar(PgCMD, PgSplit):
          'maxcnt' : 10,
          'dolock' : 1,
          'doemail' : 0,
+         'starttime' : 0,   # wall-clock start of the run, for the PBS walltime guard
+         'tardone' : 0,     # tar files dispatched so far, for the finish-rate estimate
          'cmd'  : None
       }
       self.dsids = []
@@ -105,12 +111,14 @@ class DsQuasar(PgCMD, PgSplit):
       self.dstart = 0
 
    # read in command line parameters
-   def read_parameters(self):   
+   def read_parameters(self):
       self.PGLOG['LOGFILE'] = "dsquasar.log"   # set different log file
+      self.set_help_path(__file__)
       self.dssdb_dbname()
       option = None
       argv = sys.argv[1:]
       for arg in argv:
+         if re.match(r'^-(h|-help)$', arg, re.I): self.show_usage('dsquasar')
          ms = re.match(r'^-(a|b|c|d|e|E|l|m|n|t|u|A|B|D)$', arg)
          if ms:
             arg = ms.group(1)
@@ -149,32 +157,13 @@ class DsQuasar(PgCMD, PgSplit):
                self.add_to_dsids(arg)
          else:
             self.pglog(arg + ": Value without leading option", self.LGWNEX)
-      if not (self.sopts['a'] or self.dsids):
-         print("dsquasar [-b] [(-a|-t DatasetIDs)] [-c ChangeDays] [-(e|E)] [-(B|D)] [-A ActionBits] \\")
-         print("         [-l (Y|N)] [-m ProcessCount] [-n] [-u] [-d [HostName] [TryCount]")
-         print("  Option -a - gather files of all datasets for Quasar Backup, or")
-         print("  Option -t - specify dataset IDs to backup files, wildcard % is allowed")
-         print("  Option -b - turn the background mode on; no screen display")
-         print("  Option -c - change days to work with option -A 1 to rebackup changed files")
-         print("  Option -d - turn the delay mode on for running as a batch job")
-         print("  Option -e - send an email to specialist for statistics")
-         print("  Option -E - send an email to specialist for detail statistics")
-         print("  Option -B - Quasar Backup only for the provided datasets, or")
-         print("  Option -D - Quasar Backup&Drdata; default to both B & D")
-         print("  Option -A - Action bits, 1 - create dsarch input files; 2 - build tarfiles;")
-         print("              4 - Transfer Quasar backup files; 8 - Check Backup Files;")
-         print("              16 - Dump Backup Statistics; default to 3 for 1+2 actions,")
-         print("              and the valid values are 1,2,3,4,6,7,8,16")
-         print("  Option -l - flag to check&lock dataset before backup processes, defaults to Y")
-         print("  Option -m - number of concurrent processes, defaults to 1")
-         print("  Option -n - gather and show the available file counts with no backup actions")
-         print("  Option -u - to clean up backup locks for provided dataset IDs")
-         sys.exit(0)
+      if not (self.sopts['a'] or self.dsids): self.show_usage('dsquasar')
       self.PGBACK['cmd'] = "dsquasar {}".format(' '.join(argv))
 
    # do quasar backup process
    def start_actions(self):
       self.cmdlog(self.PGBACK['cmd'])
+      self.PGBACK['starttime'] = tm()
       if self.sopts['u']:
          if self.sopts['a']: self.pglog("-u: Dataset IDs must be provided to Unlock datasets", self.LOGWRN)
          self.unlock_datasets()
@@ -331,6 +320,37 @@ class DsQuasar(PgCMD, PgSplit):
          self.backup_dataset_tarfiles(dsfiles, 'B')
          self.backup_dataset_tarfiles(dsfiles, 'D')
 
+   # size up the number of batch processes from the number of tar files that are the
+   # unit of parallel work: new infiles CINACT will create this run (estimated from the
+   # total bytes of GDEX files ready to back up, ~one tar per TARSIZE) plus existing
+   # status 'N' infile records to build (TARACT) and status 'T' tar records to transfer
+   # (BCKACT). scales one process per MPLIMIT tar files, capped at MPMAX. create-infile
+   # only (-A 1) is not a batch action (see NBACTS) so it never reaches here and stays
+   # single process.
+   def batch_process_count(self, acts):
+      tcnt = 0
+      if acts&self.CINACT:
+         sizes = [0]
+         if self.gather_dataset_files(None, False, sizes):
+            tcnt += max(1, (sizes[0] + self.TARSIZE - 1)//self.TARSIZE)
+      if acts&self.TARACT: tcnt += self.batch_tar_count('N')
+      if acts&self.BCKACT: tcnt += self.batch_tar_count('T')
+      if tcnt <= self.MPLIMIT: return 1
+      mproc = (tcnt + self.MPLIMIT - 1)//self.MPLIMIT
+      if mproc > self.MPMAX: mproc = self.MPMAX
+      return mproc
+
+   # count the bfile tar records of a given status honoring backflag/dataset scope
+   def batch_tar_count(self, status):
+      bcnd = "status = '{}'".format(status)
+      if self.PGBACK['backflag']: bcnd += " AND type = '{}'".format(self.PGBACK['backflag'])
+      if self.dsids:
+         tcnt = 0
+         for dsid in self.dsids:
+            tcnt += self.pgget('bfile', '', "dsid = '{}' AND {}".format(dsid, bcnd), self.LGWNEX)
+         return tcnt
+      return self.pgget('bfile', '', bcnd, self.LGWNEX)
+
    # start none daemon and intialize dscheck counts
    def start_dsquasar_none_daemon(self, act, fcnt = 0):
       acts = self.PGBACK['action']
@@ -338,11 +358,39 @@ class DsQuasar(PgCMD, PgSplit):
          dsid = self.dsids[0] if len(self.dsids) == 1 else ''
          if not re.match(r'^[a-z]\d{6}$', dsid): dsid = ''
          cact = ('A' if acts < 10 else '') + str(acts)
-         if act == self.STTACT or acts&self.NBACTS != acts:
-            self.set_one_boption('qoptions', '-l walltime=24:00:00', 1)
+         # only the initial delayed submit (not yet under PBS) sets qoptions for the batch
+         # job the dscheck daemon will start; size up multi-processing from the detail file
+         # count (unless an explicit -m N was given) and reserve matching PBS cpus. the
+         # invocation argv is left unchanged so a repeat submit with the same argv still
+         # blocks as a duplicate. a command-line or the PBS batch run does not set qoptions.
+         if self.PGLOG['CURBID'] < 1 and (act == self.STTACT or acts&self.NBACTS != acts):
+            # check for a duplicate command first (same lookup key init_dscheck builds);
+            # only size up multi-processing - which checks the backup files - when no
+            # matching dscheck record is on file yet, so a duplicate submit blocks without
+            # checking the backup files.
+            dargv = self.argv_to_string(sys.argv[1:], 0, "Process in Delayed Mode")
+            dargx = None
+            if len(dargv) > 100:
+               dargx = dargv[100:]
+               dargv = dargv[0:100]
+            if not self.get_dscheck("dsquasar", dargv, self.PGBACK['workdir'], self.PGLOG['CURUID'], dargx, self.LOGWRN):
+               qoptions = '-l walltime=24:00:00'
+               if acts&self.NBACTS != acts:
+                  if self.PGBACK['mproc'] < 2: self.PGBACK['mproc'] = self.batch_process_count(acts)
+                  if self.PGBACK['mproc'] > 1:
+                     qoptions += ",select=1:ncpus={0}:mem={0}gb".format(self.PGBACK['mproc'])
+               self.set_one_boption('qoptions', qoptions, 1)
          self.init_dscheck(0, '', "dsquasar", dsid, cact, self.PGBACK['workdir'],
                             self.PGLOG['CURUID'], self.bopts, self.LOGWRN)
-         if self.PGBACK['mproc'] > 1: self.PGBACK['mproc'] = 1
+         # on the PBS batch run, take the process count from the reserved ncpus in qoptions
+         # instead of recomputing it
+         if self.PGLOG['DSCHECK'] and self.PGBACK['mproc'] < 2 and acts&self.NBACTS != acts:
+            pgrec = self.pgget('dscheck', 'qoptions',
+                               "cindex = {}".format(self.PGLOG['DSCHECK']['cindex']), self.LGEREX)
+            if pgrec and pgrec['qoptions']:
+               ms = re.search(r'ncpus=(\d+)', pgrec['qoptions'])
+               if ms: self.PGBACK['mproc'] = int(ms.group(1))
+         if self.PGBACK['mproc'] > 1 and acts&self.NBACTS == acts: self.PGBACK['mproc'] = 1
       elif self.PGBACK['mproc'] > 1 and acts&self.NBACTS == acts:
          self.PGBACK['mproc'] = 1
       self.start_none_daemon('dsquasar', '', self.PGLOG['CURUID'], self.PGBACK['mproc'], 60, 1)
@@ -409,7 +457,7 @@ class DsQuasar(PgCMD, PgSplit):
       dmsg = list(bfiles)[0] if dcnt == 1 else "{} datasets".format(dcnt)
       self.pglog("{}: {} for {} ...".format(amsg, bmsg, dmsg), self.WARNLG)
       qinfo = {'backflag' : backflag, 'bid' : self.current_bid(), 'dsids' : [], 'fcnt' : 0,
-               'size' : 0, 'infiles' : [], 'instr' : '', 'qdsids' : [],
+               'size' : 0, 'infiles' : [], 'instr' : '', 'qdsids' : [], 'abids' : [],
                'dslocks' : [], 'qfcnt' : 0, 'qsize' : 0, 'qcnt' : 0}
       for dsid in bfiles:
          if len(qinfo['dsids']) == self.DSCNT: self.process_one_backup_file(qinfo, True, False)
@@ -427,6 +475,9 @@ class DsQuasar(PgCMD, PgSplit):
                self.lock_dataset(dsid, 0, self.LGEREX)
                qinfo['dslocks'].remove(dsid)
             qinfo['dsids'] = []
+      if self.PGBACK['mproc'] > 1:
+         self.check_child(None, 0, self.LOGWRN, 1)   # wait all child processes done
+         self.confirm_quasar_counts(qinfo, qinfo['abids'], "status = 'T'")   # recount confirmed backups from RDADB
       if qinfo['dslocks']:
          for dsid in qinfo['dslocks']:
             self.lock_dataset(dsid, 0, self.LGEREX)  # unlock the locked datasets
@@ -475,7 +526,7 @@ class DsQuasar(PgCMD, PgSplit):
       bmsg = self.BACKMSG[backflag]
       s = 's' if bcnt > 1 else ''
       self.pglog("{}: {} {} file{}...".format(amsg, bcnt, bmsg, s), self.WARNLG)
-      qinfo = {'backflag' : backflag, 'bid' : 0, 'dsids' : [], 'fcnt' : 0, 'size' : 0,
+      qinfo = {'backflag' : backflag, 'bid' : 0, 'dsids' : [], 'fcnt' : 0, 'size' : 0, 'abids' : [],
                'infiles' : [], 'instr' : '', 'qdsids' : [], 'qfcnt' : 0, 'qsize' : 0, 'qcnt' : 0}
       for dsid in bfiles:
          if self.PGBACK['dolock'] and self.lock_dataset(dsid, 1, self.LOGERR) < 1: continue
@@ -485,6 +536,9 @@ class DsQuasar(PgCMD, PgSplit):
             for bkey in binfo: qinfo[bkey] = binfo[bkey]
             self.process_one_backup_file(qinfo, False)
          if self.PGBACK['dolock']: self.lock_dataset(dsid, 0, self.LOGERR)
+      if self.PGBACK['mproc'] > 1:
+         self.check_child(None, 0, self.LOGWRN, 1)   # wait all child processes done
+         self.confirm_quasar_counts(qinfo, qinfo['abids'], "status = 'T'")   # recount confirmed backups from RDADB
       qcnt = qinfo['qcnt']
       if qcnt > 0:
          self.PGBACK['bckcnt'] += qcnt
@@ -522,6 +576,9 @@ class DsQuasar(PgCMD, PgSplit):
          if qinfo['size'] < self.BCKSIZE: continue
          self.transfer_quasar_tarfiles(qinfo)
       if qinfo['size'] > 0: self.transfer_quasar_tarfiles(qinfo)
+      if self.PGBACK['mproc'] > 1:
+         self.check_child(None, 0, self.LOGWRN, 1)   # wait all child processes done
+         self.confirm_quasar_counts(qinfo, list(bfiles), "status = 'A'")   # recount confirmed backups from RDADB
       qcnt = qinfo['qcnt']
       if qcnt > 0:
          self.PGBACK['bckcnt'] += qcnt
@@ -562,11 +619,67 @@ class DsQuasar(PgCMD, PgSplit):
       if self.PGBACK['pstep']: self.record_dscheck_status("F")
       self.pgexit(0)
 
+   # guard the -A 3 (Create Input&Tar, build status 'N' infiles) and -A 4 (Transfer,
+   # send status 'T' tars) PBS batch jobs against the 24-hour walltime: once past
+   # MAXRUNTIME, if the tar files still to process cannot finish within the last hour at
+   # the rate achieved so far, send a progress email and stop cleanly so the report is
+   # not lost to a PBS timeout; the remaining records are left for the next scheduled run
+   # to resume. a no-op for command-line runs, other actions, or before the cutoff.
+   def check_batch_deadline(self, qinfo):
+      act = self.PGBACK['action']
+      if self.PGLOG['CURBID'] < 1 or act not in (self.CTACTS, self.BCKACT): return
+      elapsed = tm() - self.PGBACK['starttime']
+      if elapsed < self.MAXRUNTIME: return
+      status, work = ('N', 'build') if act == self.CTACTS else ('T', 'transfer')
+      remaining = self.batch_tar_count(status)
+      if remaining < 1: return
+      done = self.PGBACK['tardone']
+      need = remaining*elapsed/done if done > 0 else elapsed
+      if need <= self.ONEHOUR: return   # enough time left to finish before the walltime
+      if self.PGBACK['mproc'] > 1: self.check_child(None, 0, self.LOGWRN, 1)   # wait all children
+      if qinfo and qinfo.get('dslocks'):
+         for dsid in qinfo['dslocks']: self.lock_dataset(dsid, 0, self.LGEREX)
+      etime = self.seconds_to_string_time(int(elapsed))
+      amsg = self.ACTMSG[act]
+      msg = "{}: Stopped after {} with {} tar file(s) still to {} - not enough time to finish before the 24-hour PBS walltime; the next scheduled run will resume".format(amsg, etime, remaining, work)
+      self.pglog(self.INDENT + msg, self.LOGACT)
+      if self.PGBACK['doemail']:
+         bmsg = self.BACKMSG[self.PGBACK['backflag']] if self.PGBACK['backflag'] else 'backup'
+         self.set_email("{}: {} - stopped early before the PBS walltime with {} {} tar file(s) remaining!".format(self.PGBACK['cmd'], amsg, remaining, bmsg), self.EMLTOP)
+         title = "dsquasar: {} stopped early ({} remaining)".format(amsg, remaining)
+         if self.PGBACK['errcnt']: title += " Error({})".format(self.PGBACK['errcnt'])
+         if self.PGLOG['DSCHECK']:
+            self.build_customized_email("dscheck", "einfo", "cindex = {}".format(self.PGLOG['DSCHECK']['cindex']), title, self.LOGWRN)
+         else:
+            self.pglog(title, self.LOGWRN|self.SNDEML)
+      if self.PGBACK['pstep']: self.record_dscheck_status("D")
+      self.pgexit(0)
+
+   # recompute the confirmed backup counts from RDADB after all child processes
+   # finished, so a multi-process summary reflects succeeded (not just started) work
+   def confirm_quasar_counts(self, qinfo, bids, dcnd):
+      qinfo['qcnt'] = qinfo['qfcnt'] = qinfo['qsize'] = 0
+      qinfo['qdsids'] = []
+      if not bids: return
+      cnd = "{} AND bid IN ({})".format(dcnd, ','.join(str(bid) for bid in bids))
+      pgrecs = self.pgmget('bfile', "dsid, dsids, data_size size, (scount + wcount) fcnt", cnd, self.LGEREX)
+      bcnt = len(pgrecs['dsid']) if pgrecs else 0
+      for i in range(bcnt):
+         qinfo['qcnt'] += 1
+         qinfo['qfcnt'] += pgrecs['fcnt'][i]
+         qinfo['qsize'] += pgrecs['size'][i]
+         dsid = pgrecs['dsid'][i]
+         if dsid not in qinfo['qdsids']: qinfo['qdsids'].append(dsid)
+         if pgrecs['dsids'][i]:
+            for did in pgrecs['dsids'][i].split(','):
+               if did and did not in qinfo['qdsids']: qinfo['qdsids'].append(did)
+
    # backup one Quasar Backup or Backup&Drdata from one or multiple inputs and,
    # reset the quasar backup dict
    def process_one_backup_file(self, qinfo, addback, keepid = False):
       ccnt = self.check_child(None, 0, self.LOGWRN, -1) if self.PGBACK['mproc'] > 1 else 0
       if self.PGSIG['QUIT']: self.quit_dsquasar(qinfo)
+      self.check_batch_deadline(qinfo)
       dsids = qinfo['dsids']
       dcnt = len(dsids)
       if dcnt == 0: return
@@ -580,6 +693,7 @@ class DsQuasar(PgCMD, PgSplit):
       else:
          bid = qinfo['bid']
          qfile = qinfo['bfile']
+      if 'abids' in qinfo: qinfo['abids'].append(bid)   # track dispatched bids for confirmed recount
       stat = 1
       if self.PGBACK['action'] > self.CINACT:
          s = 's' if fcnt > 1 else ''
@@ -601,7 +715,7 @@ class DsQuasar(PgCMD, PgSplit):
                   for infile in qinfo['infiles']: self.delete_local_file(infile)
                elif re.search(r'file backed up to', self.PGLOG['SYSERR']):
                   if self.pgdel('bfile', f"bid = {bid}"):
-                     self.pglog(f"{dsid}-{qfile}: backup tarfile deleted for duplicattion", self.DTLACT)
+                     self.pglog(f"{dsid}-{qfile}: backup tarfile deleted for duplication", self.DTLACT)
                   for infile in qinfo['infiles']: self.delete_local_file(infile)
                sys.exit(0)  # stop child process
             else:
@@ -614,7 +728,7 @@ class DsQuasar(PgCMD, PgSplit):
                for infile in qinfo['infiles']: self.delete_local_file(infile)
             elif re.search(r'file backed up to', self.PGLOG['SYSERR']):
                if self.pgdel('bfile', f"bid = {bid}"):
-                  self.pglog(f"{dsid}-{qfile}: backup tarfile deleted for duplicattion", self.DTLACT)
+                  self.pglog(f"{dsid}-{qfile}: backup tarfile deleted for duplication", self.DTLACT)
                for infile in qinfo['infiles']: self.delete_local_file(infile)
             else:
                self.PGBACK['errcnt'] += 1
@@ -623,12 +737,13 @@ class DsQuasar(PgCMD, PgSplit):
       if stat:
          # reset qinfo after quasar backup
          qinfo['qcnt'] += 1
+         self.PGBACK['tardone'] += 1   # cumulative dispatched tars for the walltime finish-rate
          qinfo['qfcnt'] += fcnt
          qinfo['qsize'] += fsize
          for dsid in dsids:
             if dsid not in qinfo['qdsids']: qinfo['qdsids'].append(dsid)
       if addback:
-         lastdsid = dsids.pop() if keepid else None         
+         lastdsid = dsids.pop() if keepid else None
          for dsid in dsids:  # unlock all datasets but the last one
             if dsid in qinfo['dslocks']:
                self.lock_dataset(dsid, 0, self.LGEREX)
@@ -645,6 +760,7 @@ class DsQuasar(PgCMD, PgSplit):
    def transfer_quasar_tarfiles(self, qinfo):
       ccnt = self.check_child(None, 0, self.LOGWRN, -1) if self.PGBACK['mproc'] > 1 else 0
       if self.PGSIG['QUIT']: self.quit_dsquasar(qinfo)
+      self.check_batch_deadline(qinfo)
       # prepare for backup one tar file
       dsids = qinfo['dsids']
       bids = qinfo['bids']
@@ -675,10 +791,10 @@ class DsQuasar(PgCMD, PgSplit):
       dstat = bstat = -1
       if ccnt == 0 or self.PGSIG['PPID'] > 1:
          if bflg == 'D':
-            dstat = self.quasar_multiple_trasnfer(tofiles, fromfiles, 'gdex-quasar-drdata', 'gdex-glade', self.ERRACT)
+            dstat = self.quasar_multiple_transfer(tofiles, fromfiles, 'gdex-quasar-drdata', 'gdex-glade', self.ERRACT)
             if not dstat: self.pglog("Error Quaser Drdata for " + fmsg, self.ERRACT|self.LOGERR)
-         bstat = self.quasar_multiple_trasnfer(tofiles, fromfiles, 'gdex-quasar', 'gdex-glade', self.ERRACT)
-         if not dstat: self.pglog("Error Quaser Backup for " + fmsg, self.ERRACT|self.LOGERR)
+         bstat = self.quasar_multiple_transfer(tofiles, fromfiles, 'gdex-quasar', 'gdex-glade', self.ERRACT)
+         if not bstat: self.pglog("Error Quaser Backup for " + fmsg, self.ERRACT|self.LOGERR)
          if dstat == self.FINISH: dstat = self.check_globus_finished(tofiles[0], 'gdex-quasar-drdata', self.ERRACT|self.NOWAIT)
          if bstat == self.FINISH: bstat = self.check_globus_finished(tofiles[0], 'gdex-quasar', self.ERRACT|self.NOWAIT)
          if dstat and bstat:
@@ -698,6 +814,7 @@ class DsQuasar(PgCMD, PgSplit):
       if dstat and bstat:
          # reset qinfo after quasar backup
          qinfo['qcnt'] += bcnt
+         self.PGBACK['tardone'] += bcnt   # cumulative transferred tars for the walltime finish-rate
          qinfo['qfcnt'] += fcnt
          qinfo['qsize'] += fsize
          for dsid in dsids:
@@ -777,7 +894,7 @@ class DsQuasar(PgCMD, PgSplit):
       return {'wcount' : 0, 'wrecs' : [], 'scount' : 0, 'srecs' : []}
 
    # get all files need to be backed up for a given dataset ID
-   def get_dataset_files(self, dsid, dsfiles, backflag, logact = 0):
+   def get_dataset_files(self, dsid, dsfiles, backflag, logact = 0, sizes = None):
       dcnd = "dsid = '{}'".format(dsid)
       fopt = self.get_backup_options(dsid, dcnd)
       if not fopt:
@@ -817,6 +934,7 @@ class DsQuasar(PgCMD, PgSplit):
                self.pglog("{}-G{}: Found {} Files ({})".format(dsid, gidx, rcnt, rmsg), self.LOGWRN)
    
       fcnt = cnts['B'][0] + cnts['D'][0]
+      if sizes is not None: sizes[0] += cnts['B'][1] + cnts['D'][1]
       if dsfiles:
          cmsg = 'Changed ' if self.PGBACK['chgdays'] else ''
          for bflg in dsfiles:
@@ -836,7 +954,6 @@ class DsQuasar(PgCMD, PgSplit):
 
    # get web/saved files to backup for a given group condition and file option
    def get_group_files(self, dsid, gcnd, dcnd, bfiles, fopt, infiles, cnts):
-      rcnt = 0
       if self.PGBACK['chgdays'] > 0:
          return self.get_group_changed_files(dsid, gcnd, dcnd, bfiles, fopt, infiles, cnts)
       else:
@@ -920,7 +1037,7 @@ class DsQuasar(PgCMD, PgSplit):
       fname = pgrec[tname]
       chome = self.PGLOG['DECSHOME'] if cate == 's' else self.PGLOG['DSDHOME']
       if pgrec['locflag'] == 'O':
-         bucket = 'gdex-decsdata' if cate == 's' else self.PGLOC['OBJCTBKT']
+         bucket = 'gdex-decsdata' if cate == 's' else self.PGLOG['OBJCTBKT']
          ofile = self.join_paths(dsid, fname)
          info = self.check_object_file(ofile, bucket, 1, self.ERRACT)
          if not info: return False
@@ -982,28 +1099,27 @@ class DsQuasar(PgCMD, PgSplit):
    def get_backup_options(self, dsid, dcnd):
       fopt = 0
       bcnd = "= 0" if self.PGBACK['chgdays'] < 1 else "> 0"
-   #   if self.pgget('sfile', 'sid', "{} AND bid {}".format(dcnd, bcnd), self.LGWNEX): fopt |= self.SOPT
       if self.pgget_wfile(dsid, 'wid', "bid " + bcnd, self.LGWNEX): fopt |= self.WOPT
       return fopt
 
    # gather all available dataset ids to backup data files
-   def gather_dataset_files(self, dsfiles, unlock = True):
+   def gather_dataset_files(self, dsfiles, unlock = True, sizes = None):
       fcnt = 0
+      # warn per named dataset with no files; stay silent when scanning all datasets
+      logact = self.LOGWRN if self.dsids else 0
       if self.dsids:
+         pgrecs = []
          for dsid in self.dsids:
-            dcnd = "dsid = '{}'".format(dsid)
-            pgrec = self.pgget("dataset", "dsid, backflag, pid", dcnd, self.LGWNEX)
-            if pgrec:
-               if unlock and pgrec['pid'] and self.lock_dataset(dsid, 0, self.LOGACT) < 1: continue
-               fcnt += self.get_dataset_files(dsid, dsfiles, pgrec['backflag'], self.LOGWRN)
+            pgrec = self.pgget("dataset", "dsid, backflag, pid", "dsid = '{}'".format(dsid), self.LGWNEX)
+            if pgrec: pgrecs.append(pgrec)
       else:
-         dcnd = "ORDER BY dsid"
-         pgrecs = self.pgmget("dataset", "dsid, backflag, pid", dcnd, self.LGWNEX)
-         dcnt = len(pgrecs['dsid']) if pgrecs else 0
-         for i in range(dcnt):
-            dsid = pgrecs['dsid'][i]
-            if unlock and pgrecs['pid'][i] and self.lock_dataset(dsid, 0, self.LOGACT) < 1: continue
-            fcnt += self.get_dataset_files(dsid, dsfiles, pgrecs['backflag'][i])
+         mrecs = self.pgmget("dataset", "dsid, backflag, pid", "ORDER BY dsid", self.LGWNEX)
+         dcnt = len(mrecs['dsid']) if mrecs else 0
+         pgrecs = [self.onerecord(mrecs, i) for i in range(dcnt)]
+      for pgrec in pgrecs:
+         dsid = pgrec['dsid']
+         if unlock and pgrec['pid'] and self.lock_dataset(dsid, 0, self.LOGACT) < 1: continue
+         fcnt += self.get_dataset_files(dsid, dsfiles, pgrec['backflag'], logact, sizes)
       if dsfiles:
          s = 's' if fcnt > 1 else ''
          bmsg = self.BACKMSG[self.PGBACK['backflag']] if self.PGBACK['backflag'] else 'backup' 
@@ -1013,28 +1129,21 @@ class DsQuasar(PgCMD, PgSplit):
 
    # gather all available backup file records with status N to backup
    def gather_dataset_infiles(self, dsfiles):
-      icnt = fcnt = cnt = 0
+      icnt = fcnt = 0
       bcnd = "status = 'N'"
-      ocnd = " ORDER BY dsid"
       flds = "bid, dsid, bfile, type, data_size size, (scount + wcount) fcnt, dsids, note"
       if self.PGBACK['backflag']: bcnd += " AND type = '{}'".format(self.PGBACK['backflag'])
       if self.dsids:
-         for dsid in self.dsids:
-            dcnd = "dsid = '{}' AND {}".format(dsid, bcnd)
-            pgrecs = self.pgmget("bfile", flds, dcnd, self.LGWNEX)
-            bcnt = len(pgrecs['bid']) if pgrecs else 0
-            for i in range(bcnt):
-               cnt = self.get_backup_infile(dsfiles, self.onerecord(pgrecs, i))
-               if cnt:
-                  icnt +=1
-                  fcnt += cnt
+         cnds = ["dsid = '{}' AND {}".format(dsid, bcnd) for dsid in self.dsids]
       else:
-         pgrecs = self.pgmget("bfile", flds, bcnd+ocnd, self.LGWNEX)
+         cnds = [bcnd + " ORDER BY dsid"]
+      for dcnd in cnds:
+         pgrecs = self.pgmget("bfile", flds, dcnd, self.LGWNEX)
          bcnt = len(pgrecs['bid']) if pgrecs else 0
          for i in range(bcnt):
             cnt = self.get_backup_infile(dsfiles, self.onerecord(pgrecs, i))
             if cnt:
-               icnt +=1
+               icnt += 1
                fcnt += cnt
       if dsfiles:
          s = 's' if icnt > 1 else ''
@@ -1045,28 +1154,21 @@ class DsQuasar(PgCMD, PgSplit):
    # gather all available backup file records with status T to backup
    #
    def gather_dataset_tarfiles(self, dsfiles):
-      tcnt = fcnt = cnt = 0
+      tcnt = fcnt = 0
       bcnd = "status = 'T'"
-      ocnd = " ORDER BY dsid"
       flds = "bid, dsid, bfile, type, data_size size, (scount + wcount) fcnt, note"
       if self.PGBACK['backflag']: bcnd += " AND type = '{}'".format(self.PGBACK['backflag'])
       if self.dsids:
-         for dsid in self.dsids:
-            dcnd = "dsid = '{}' AND {}".format(dsid, bcnd)
-            pgrecs = self.pgmget("bfile", flds, dcnd, self.LGWNEX)
-            bcnt = len(pgrecs['bid']) if pgrecs else 0
-            for i in range(bcnt):
-               cnt = self.get_backup_tarfile(dsfiles, self.onerecord(pgrecs, i))
-               if cnt:
-                  tcnt +=1
-                  fcnt += cnt
+         cnds = ["dsid = '{}' AND {}".format(dsid, bcnd) for dsid in self.dsids]
       else:
-         pgrecs = self.pgmget("bfile", flds, bcnd+ocnd, self.LGWNEX)
+         cnds = [bcnd + " ORDER BY dsid"]
+      for dcnd in cnds:
+         pgrecs = self.pgmget("bfile", flds, dcnd, self.LGWNEX)
          bcnt = len(pgrecs['bid']) if pgrecs else 0
          for i in range(bcnt):
             cnt = self.get_backup_tarfile(dsfiles, self.onerecord(pgrecs, i))
             if cnt:
-               tcnt +=1
+               tcnt += 1
                fcnt += cnt
       s = 's' if tcnt > 1 else ''
       bmsg = self.BACKMSG[self.PGBACK['backflag']] if self.PGBACK['backflag'] else 'backup' 
@@ -1092,7 +1194,6 @@ class DsQuasar(PgCMD, PgSplit):
       if self.PGBACK['action']&self.CINACT: ret = fcnt
       if not dsfiles: return ret
       dsids = [dsid]
-   #   if not (pgrec['note'] or self.rebuild_file_note(pgrec, dsids)): return 0
       if not pgrec['note']:
          return self.pglog(bfile + ": Miss note for N backup file", self.ERRACT)
       infiles = []
@@ -1147,7 +1248,6 @@ class DsQuasar(PgCMD, PgSplit):
       if self.PGBACK['action']&self.CINACT: ret = fcnt
       if not dsfiles: return ret
       dsids = [dsid]
-   #   if not (pgrec['note'] or self.rebuild_file_note(pgrec, dsids)): return 0
       if not pgrec['note']:
          return self.pglog(bfile + ": Miss note for T backup file", self.ERRACT)
       ftype = None
@@ -1347,6 +1447,9 @@ class DsQuasar(PgCMD, PgSplit):
          binfo = bfiles[bid]
          for bkey in binfo: qinfo[bkey] = binfo[bkey]
          self.process_one_quasar_pathfile(qinfo)
+      if self.PGBACK['mproc'] > 1:
+         self.check_child(None, 0, self.LOGWRN, 1)   # wait all child processes done
+         self.confirm_quasar_counts(qinfo, list(bfiles), "bfile LIKE 'G%/%.tar'")   # recount confirmed renames from RDADB
       qcnt = qinfo['qcnt']
       if qcnt > 0:
          s = 's' if qcnt > 1 else ''
@@ -1671,43 +1774,6 @@ class DsQuasar(PgCMD, PgSplit):
       dsfiles[pgrec['type']][pgrec['bid']] = {'note' : pgrec['note'], 'fcnt' : pgrec['fcnt'], 'size' : pgrec['size'],
                                               'date' : pgrec['date_modified'], 'time' : pgrec['time_modified']}
 
-   # rebuild missed file note 
-   def rebuild_file_note(self, pgrec, dsids):
-      bid = pgrec['bid']
-      fields = 'file, dsid, type, data_size, checksum'
-      bfiles = {dsids[0] : {'S' : [], 'W' : []}}
-      pgrecs = self.pgmget('sfile', 's' + fields, "bid = {} ORDER BY dsid, sfile".format(bid), self.LGEREX)
-      cnt = len(pgrecs['sfile']) if pgrecs else 0
-      for i in range(cnt):
-         dsid = pgrecs['dsid'][i]
-         if dsid not in dsids:
-            dsids.append(dsid)
-            bfiles[dsid] = {'S' : [], 'W' : []}
-         bfiles[dsid]['S'].append([pgrecs['sfile'][i], pgrecs['type'][i], pgrecs['data_size'][i], pgrecs['checksum'][i]])
-      pgrecs = self.pgmget('wfile', 'w' + fields, "bid = {} ORDER BY dsid, wfile".format(bid), self.LGEREX)
-      cnt = len(pgrecs['wfile']) if pgrecs else 0
-      for i in range(cnt):
-         dsid = pgrecs['dsid'][i]
-         if dsid not in dsids:
-            dsids.append(dsid)
-            bfiles[dsid] = {'S' : [], 'W' : []}
-         bfiles[dsid]['W'].append([pgrecs['wfile'][i], pgrecs['type'][i], pgrecs['data_size'][i], pgrecs['checksum'][i]])
-      # create note
-      note = ''
-      for dsid in bfiles:
-         for ftype in bfiles[dsid]:
-            files = bfiles[dsid][ftype]
-            if not files: continue
-            note += "<{}_{}_{}.txt>\n".format(dsid, ftype, bid)
-            note += "DS<=>{}\n{}F<:>{}T<:>SZ<:>MC<:>\n".format(dsid, ftype, ftype)
-            for flist in files:
-               note += "{}<:>{}<:>{}<:>{}<:>\n".format(flist[0], flist[1], flist[2], flist[3])
-      if note:
-         self.pgexec("UPDATE bfile SET note = '{}' WHERE bid = {}".format(note, bid), self.LGEREX)
-         pgrec['note'] = note
-         return 1
-      return 0
-
    # dump statistics for GDEX files available to create input files
    def dump_dataset_files(self, dsfiles, backflag):
       bfiles = dsfiles[backflag]
@@ -1922,7 +1988,7 @@ class DsQuasar(PgCMD, PgSplit):
             pcnt += 1
       return pcnt
 
-# main function to excecute this script
+# main function to execute this script
 def main():
    object = DsQuasar()
    object.read_parameters()

rda_python_dsquasar-3.0.3/src/rda_python_dsquasar/dsquasar.usg

@@ -0,0 +1,128 @@
+
+ Name: dsquasar - back up RDA/GDEX data archives onto the Quasar backup servers
+
+ Check the Backup flags set in dataset.backflag and/or dsgroup.backflag of RDADB
+ to gather Web and Saved files, tar them into larger (>= 2GB) backup files, and
+ back them up onto the NCAR GDEX Quasar Globus end points, for a Backup copy
+ (gdex-quasar) and/or a disaster-recovery Drdata copy (gdex-quasar-drdata).  The
+ file tarring and transferring is performed through the utility program dsarch.
+
+ Usage: dsquasar [-b] [(-a | -t DatasetIDs)] [-c ChangeDays] [-(e|E)]     \
+                 [-(B|D)] [-A ActionBits] [-l (Y|N)] [-m ProcessCount]    \
+                 [-n] [-u] [-d [HostName] [TryCount]] [-h]
+
+   Either -a or -t must be given (usage is displayed otherwise).  dsquasar runs
+   from the Quasar backup work directory; it changes there automatically.
+
+
+ DATASET SELECTION
+
+   -a     Gather files of all datasets that have a Backup flag set.
+
+   -t  DatasetIDs
+          One or more dataset IDs whose files are backed up.  The SQL wildcard
+          '%' is allowed and is expanded against the dataset table.  Ignored,
+          with a warning, when -a is also given.
+
+
+ BACKUP SCOPE
+
+   -B     Quasar Backup only (endpoint gdex-quasar) for the given datasets.
+
+   -D     Quasar Backup & Drdata (gdex-quasar and gdex-quasar-drdata).
+
+          -B and -D are mutually exclusive; the default (neither given) backs up
+          to both the Backup and the Drdata endpoints.
+
+   -c  ChangeDays
+          Used with -A 1 to re-back up files that changed within the given
+          number of days.  Must be greater than 0.
+
+
+ ACTIONS
+
+   -A  ActionBits
+          Bit flags selecting the actions to perform; the valid values are
+          1, 2, 3, 4, 6, 7, 8, 16 and it defaults to 3 (actions 1 + 2):
+
+             1  - create the dsarch input Save/Web file lists and add the
+                  backup file (bfile) placeholder records (status 'N').
+             2  - build tar files (each >= 5GB where possible) from the Saved/
+                  Web files listed in the input files (status 'N' -> 'T').
+             4  - transfer the built tar files (batched to >= 90GB per Globus
+                  transfer) to the Quasar backup end points (status 'T' -> 'A').
+             8  - check file sizes between RDADB and the Quasar servers, and
+                  reset status to 'N' when they differ.
+            16  - dump backup statistics for files backed up and files that are
+                  ready to be backed up.
+
+          Combined values: 3 = 1+2, 6 = 2+4, 7 = 1+2+4.
+
+
+ RUN CONTROL
+
+   -b     Turn on background mode; no screen display.
+
+   -d  [HostName] [TryCount]
+          Turn on delay mode to run as a batch job.  A dscheck record is created
+          and the dscheck daemon later submits the job to PBS.  On the initial
+          submit dsquasar sizes multi-processing from the detail file count and
+          reserves matching PBS cpus (walltime + select=1:ncpus=N:mem=Ngb) in the
+          dscheck qoptions; the invocation arguments are left unchanged so a
+          repeat submit with the same arguments still blocks as a duplicate.
+
+   -l  (Y|N)
+          Whether to check and lock a dataset before backing it up.  Defaults
+          to Y.
+
+   -m  ProcessCount
+          Number of concurrent child processes.  Defaults to 1.  In delay mode
+          the process count travels through the reserved PBS cpus rather than
+          the command line.
+
+   -n     Gather and show the available file counts only; no backup actions are
+          performed.
+
+   -u     Clean up (unlock) backup locks for the given dataset IDs.  Dataset IDs
+          must be provided (not valid with -a).
+
+
+ EMAIL
+
+   -e     Send an email to the specialist with backup statistics.
+
+   -E     Send an email to the specialist with detailed backup statistics.
+
+
+ MISCELLANEOUS
+
+   -h     Display this help document.
+
+
+ EXAMPLES
+
+   1. Back up (create input files and tar files for) all flagged datasets:
+
+         dsquasar -a -A 3
+
+   2. Transfer built tar files of one dataset to both Quasar end points:
+
+         dsquasar -t d123456 -A 4
+
+   3. Show the counts of files ready to back up for a dataset, without acting:
+
+         dsquasar -t d123456 -n
+
+   4. Dump backup statistics for all datasets by email:
+
+         dsquasar -a -A 16 -e
+
+   5. Queue a delayed PBS batch job to create input files and tar files for all
+      datasets in the background (as run from cron); multi-processing is sized
+      automatically from the file count:
+
+         dsquasar -a -A 3 -e -b -d PBS
+
+   6. Unlock backup locks left on datasets:
+
+         dsquasar -t d123456 -u

{rda_python_dsquasar-2.0.12 → rda_python_dsquasar-3.0.3}/src/rda_python_dsquasar/dstacc.py

@@ -81,7 +81,7 @@ class DsTACC(PgCMD, PgSplit):
       self.ONESIZE = 20*self.PGLOG['ONEGBS']  # 20GB, minimal file size to tar a single file 
       self.TFCOUNT = 1000000      # if file count is greater, use self.MINSIZE for tar file
       self.PGBACK = {
-         'workdir' : "{}/{}/tacc_backup".format(self.PGLOG['GDEXWORK'], self.PGLOG['GDEXUSER']),
+         'workdir' : "{}/{}/tacc_backup".format(self.PGLOG['GDEXWORK'], self.PGLOG['COMMONUSER']),
          'mproc' : 1,
          'action' : self.CTACTS,
          'chgdays' : 0,
@@ -857,7 +857,7 @@ class DsTACC(PgCMD, PgSplit):
       fname = pgrec[tname]
       chome = self.PGLOG['DECSHOME'] if cate == 's' else self.PGLOG['DSDHOME']
       if pgrec['locflag'] == 'O':
-         bucket = 'gdex-decsdata' if cate == 's' else self.PGLOC['OBJCTBKT']
+         bucket = 'gdex-decsdata' if cate == 's' else self.PGLOG['OBJCTBKT']
          ofile = self.join_paths(dsid, fname)
          info = self.check_object_file(ofile, bucket, 1, self.ERRACT)
          if not info: return False

{rda_python_dsquasar-2.0.12 → rda_python_dsquasar-3.0.3}/src/rda_python_dsquasar/taccrec.py

@@ -19,7 +19,7 @@ def get_tar_summary_and_details(member_list_path):
         with open(member_list_path, 'r') as f:
             for idx, line in enumerate(f):
                 line = line.strip()
-                if not line or line.startswith('-rw') is False:
+                if not line:
                     continue
                 parts = line.split()
                 if len(parts) < 6:

rda_python_dsquasar-3.0.3/src/rda_python_dsquasar.egg-info/PKG-INFO

@@ -0,0 +1,87 @@
+Metadata-Version: 2.4
+Name: rda_python_dsquasar
+Version: 3.0.3
+Summary: RDA Python package to backup and recover RDA data archives to and from GLOBUS Quasar backup server
+Author-email: Zaihua Ji <zji@ucar.edu>
+Project-URL: Homepage, https://github.com/NCAR/rda-python-dsquasar
+Classifier: Programming Language :: Python :: 3
+Classifier: License :: OSI Approved :: MIT License
+Classifier: Operating System :: OS Independent
+Classifier: Development Status :: 5 - Production/Stable
+Requires-Python: >=3.7
+Description-Content-Type: text/markdown
+License-File: LICENSE
+Requires-Dist: rda_python_common
+Requires-Dist: rda_python_dsarch
+Dynamic: license-file
+
+# rda_python_dsquasar
+
+RDA Python package to back up and recover RDA data archives to and from the
+GLOBUS Quasar backup server.
+
+## Programs
+
+The package installs the following command-line utilities, all running as the
+current user:
+
+- `dsquasar` — back up and restore RDA dataset archives via Quasar
+- `tacctar` — create tar bundles for Quasar archival
+- `taccrec` — recover content from Quasar tar bundles
+
+Run any program with `--help` (or `-h`) for full usage details.
+
+## Environment setup
+
+Create a Python environment first; package installs in the next section run
+inside whichever environment you activate here.
+
+### Option A — Python venv (DECS machines)
+
+```bash
+python3 -m venv $ENVHOME          # e.g. /glade/u/home/gdexdata/gdexmsenv
+source $ENVHOME/bin/activate
+```
+
+### Option B — Conda (DAV/Casper)
+
+```bash
+conda create --prefix $ENVHOME python=3.12   # e.g. /glade/work/gdexdata/conda-envs/pg-gdex
+conda activate $ENVHOME
+```
+
+## Installing rda-python-dsquasar
+
+Pick whichever install mode fits your workflow.  All variants pull in the
+transitive dependencies (`rda_python_common`, `rda_python_dsarch`)
+automatically.
+
+For local development, clone this repo alongside your project and install it
+in editable mode so that changes are picked up without re-installing:
+
+```bash
+git clone https://github.com/NCAR/rda-python-dsquasar.git
+cd rda-python-dsquasar
+pip install -e .
+```
+
+To test a specific branch (e.g. an in-progress feature or fix branch), pass
+`-b/--branch` to `git clone`:
+
+```bash
+git clone -b <branch-name> https://github.com/NCAR/rda-python-dsquasar.git
+cd rda-python-dsquasar
+pip install -e .
+```
+
+For a regular (non-editable) install from a checkout:
+
+```bash
+pip install /path/to/rda-python-dsquasar
+```
+
+For a production install on a system that uses the published distribution:
+
+```bash
+pip install rda_python_dsquasar
+```

{rda_python_dsquasar-2.0.12 → rda_python_dsquasar-3.0.3}/src/rda_python_dsquasar.egg-info/SOURCES.txt

@@ -1,9 +1,11 @@
 LICENSE
+MANIFEST.in
 README.md
 pyproject.toml
 src/rda_python_dsquasar/__init__.py
 src/rda_python_dsquasar/ds_quasar.py
 src/rda_python_dsquasar/dsquasar.py
+src/rda_python_dsquasar/dsquasar.usg
 src/rda_python_dsquasar/dstacc.py
 src/rda_python_dsquasar/taccrec.py
 src/rda_python_dsquasar/tacctar.py

rda_python_dsquasar-2.0.12/PKG-INFO

@@ -1,18 +0,0 @@
-Metadata-Version: 2.4
-Name: rda_python_dsquasar
-Version: 2.0.12
-Summary: RDA Python package to backup and recover RDA data archives to and from GLOBUS Quasar backup server
-Author-email: Zaihua Ji <zji@ucar.edu>
-Project-URL: Homepage, https://github.com/NCAR/rda-python-dsquasar
-Classifier: Programming Language :: Python :: 3
-Classifier: License :: OSI Approved :: MIT License
-Classifier: Operating System :: OS Independent
-Classifier: Development Status :: 5 - Production/Stable
-Requires-Python: >=3.7
-Description-Content-Type: text/markdown
-License-File: LICENSE
-Requires-Dist: rda_python_common
-Requires-Dist: rda_python_dsarch
-Dynamic: license-file
-
-RDA python package to backup RDA dataset data onto Globus Quasar Server.

rda_python_dsquasar-2.0.12/README.md

@@ -1 +0,0 @@
-RDA python package to backup RDA dataset data onto Globus Quasar Server.

rda_python_dsquasar-2.0.12/src/rda_python_dsquasar.egg-info/PKG-INFO

@@ -1,18 +0,0 @@
-Metadata-Version: 2.4
-Name: rda_python_dsquasar
-Version: 2.0.12
-Summary: RDA Python package to backup and recover RDA data archives to and from GLOBUS Quasar backup server
-Author-email: Zaihua Ji <zji@ucar.edu>
-Project-URL: Homepage, https://github.com/NCAR/rda-python-dsquasar
-Classifier: Programming Language :: Python :: 3
-Classifier: License :: OSI Approved :: MIT License
-Classifier: Operating System :: OS Independent
-Classifier: Development Status :: 5 - Production/Stable
-Requires-Python: >=3.7
-Description-Content-Type: text/markdown
-License-File: LICENSE
-Requires-Dist: rda_python_common
-Requires-Dist: rda_python_dsarch
-Dynamic: license-file
-
-RDA python package to backup RDA dataset data onto Globus Quasar Server.