Fixed issues with the db and logs not being recovered in different ru…

…n scenearios

* Update VERSION

* added process_jobs_to_submit fixes runtime issue

* Added few todos that requires to change critical stuff

* Fix runtime issue, recovery now shows a better output

* Improved package test

* Improved package test

* Closing file descriptors on reconnect.

* Fixes an issue with logs with this sequence of commands: `create + run + create + recovery + run`

* Intregation test added, fixed few bugs

* Fixed DB test

* Now the db test also checks the stat and job_names in the filesystem

* feedback

* Changed test name,
Added generic functions in the test to expand it with other run options

* Added few todos

* disabled the check exit code as it is not working on the pipeline and not important right now

* Adding more chunks to the wrapper success

* More detailed test

* Add CI and dependabot GitHub actions (#2021)

* add CI and dependabot gh actions

* disable linting temporarily

(cherry picked from commit 8a1b2df)

* Improved test output in case of failure

* more info

* more info

* Test adding +x

* added chmod after sending

* chmod added

---------

Co-authored-by: dbeltran <daniel.beltran@bsc.es>
Co-authored-by: Luiggi Tenorio <luiggibit@gmail.com>

---------

Co-authored-by: dbeltran <daniel.beltran@bsc.es>
Co-authored-by: Luiggi Tenorio <luiggibit@gmail.com>

.github/dependabot.yml

@@ -0,0 +1,25 @@
+# To get started with Dependabot version updates, you'll need to specify which
+# package ecosystems to update and where the package manifests are located.
+# Please see the documentation for all configuration options:
+# https://docs.github.com/code-security/dependabot/dependabot-version-updates/configuration-options-for-the-dependabot.yml-file
+
+version: 2
+updates:
+
+  # Maintain dependencies for GitHub Actions
+  - package-ecosystem: "github-actions"
+    # Workflow files stored in the default location of `.github/workflows`. (You don't need to specify `/.github/workflows` for `directory`. You can use `directory: "/"`.)
+    directory: "/"
+    schedule:
+      interval: "daily"
+    labels:
+      - 'dependencies'
+
+  # Maintain dependencies for PyPI
+  - package-ecosystem: "pip" # See documentation for possible values
+    directory: "/" # Location of package manifests
+    schedule:
+      interval: "daily"
+    labels:
+      - 'dependencies'
+

.github/workflows/ci.yaml

@@ -0,0 +1,95 @@
+name: CI
+
+on:
+  pull_request:
+  workflow_dispatch:
+  push:
+    branches:
+      - main
+
+jobs:
+  # lint: # Turn on when linting issues are resolved
+  #   runs-on: ubuntu-latest
+  #   timeout-minutes: 2
+
+  #   steps:
+  #     - name: Checkout code
+  #       uses: actions/checkout@v4
+
+  #     - name: Set up Python
+  #       uses: actions/setup-python@v5
+  #       with:
+  #         python-version: "3.9"
+
+  #     - name: Install dependencies
+  #       run: |
+  #         python -m pip install --upgrade pip
+  #         pip install -e .[all]
+
+  #     - name: Lint code
+  #       run: |
+  #         ruff check .
+
+  test:
+    # needs: lint
+    runs-on: ubuntu-latest
+    timeout-minutes: 5
+
+    strategy:
+      matrix:
+        python-version: ["3.9", "3.10", "3.11", "3.12"]
+
+    steps:
+      - name: Checkout code
+        uses: actions/checkout@v4
+
+      - name: Set up Python
+        uses: actions/setup-python@v5
+        with:
+          python-version: ${{ matrix.python-version }}
+
+      - name: Install system dependencies
+        run: sudo apt-get install -y graphviz rsync curl
+
+      - name: Install dependencies
+        run: |
+          python -m pip install --upgrade pip
+          pip install -e .[all]
+
+      - name: Run tests
+        run: |
+          pytest
+
+      - name: Coverage report
+        run: |
+          coverage xml
+          coverage report
+
+      - name: Upload coverage artifact
+        uses: actions/upload-artifact@v4
+        with:
+          name: coverage_${{ matrix.os }}_py-${{ matrix.python-version }}
+          path: coverage.xml
+          retention-days: 7
+
+  coverage:
+    needs: test
+    runs-on: ubuntu-latest
+    timeout-minutes: 2
+    steps:
+      - name: Checkout
+        uses: actions/checkout@v4
+
+      - name: Download coverage artifacts
+        uses: actions/download-artifact@v4
+
+      - name: Codecov upload
+        uses: codecov/codecov-action@v5
+        with:
+          name: ${{ github.workflow }}
+          flags: fast-tests
+          fail_ci_if_error: true
+          verbose: true
+          # Token not required for public repos, but avoids upload failure due
+          # to rate-limiting (but not for PRs opened from forks)
+          token: ${{ secrets.CODECOV_TOKEN }}

VERSION

@@ -1 +1 @@
-4.1.11
+4.1.12

autosubmit/autosubmit.py

@@ -1743,27 +1743,6 @@ def generate_scripts_andor_wrappers(as_conf, job_list, jobs_filtered, packages_p
         for job in job_list.get_job_list():
             job.status = Status.WAITING
 
-    @staticmethod
-    def terminate(all_threads):
-        # Closing threads on Ctrl+C
-        Log.info(
-            "Looking for active threads before closing Autosubmit. Ending the program before these threads finish may result in unexpected behavior. This procedure will last until all threads have finished or the program has waited for more than 30 seconds.")
-        timeout = 0
-        active_threads = True
-        while active_threads and timeout <= 60:
-            active_threads = False
-            for thread in all_threads:
-                if "JOB_" in thread.name:
-                    if thread.is_alive():
-                        active_threads = True
-                        Log.info("{0} is still retrieving outputs, time remaining is {1} seconds.".format(
-                            thread.name, 60 - timeout))
-                        break
-            if active_threads:
-                sleep(10)
-                timeout += 10
-
-
     @staticmethod
     def manage_wrapper_job(as_conf, job_list, platform, wrapper_id, save=False):
         check_wrapper_jobs_sleeptime = as_conf.get_wrapper_check_time()
@@ -2144,7 +2123,12 @@ def get_iteration_info(as_conf,job_list):
         Log.debug("Sleep: {0}", safetysleeptime)
         Log.debug("Number of retrials: {0}", default_retrials)
         return total_jobs, safetysleeptime, default_retrials, check_wrapper_jobs_sleeptime
-
+
+    @staticmethod
+    def check_logs_status(job_list, as_conf, new_run):
+        for job in job_list.get_completed_failed_without_logs():
+            job_list.update_log_status(job, as_conf, new_run)
+
     @staticmethod
     def run_experiment(expid, notransitive=False, start_time=None, start_after=None, run_only_members=None, profile=False):
         """
@@ -2203,6 +2187,7 @@ def run_experiment(expid, notransitive=False, start_time=None, start_after=None,
 
                 max_recovery_retrials = as_conf.experiment_data.get("CONFIG",{}).get("RECOVERY_RETRIALS",3650)  # (72h - 122h )
                 recovery_retrials = 0
+                Autosubmit.check_logs_status(job_list, as_conf, new_run=True)
                 while job_list.get_active():
                     for platform in platforms_to_test:  # Send keep_alive signal
                         platform.work_event.set()
@@ -2211,7 +2196,7 @@ def run_experiment(expid, notransitive=False, start_time=None, start_after=None,
                     did_run = True
                     try:
                         if Autosubmit.exit:
-                            Autosubmit.terminate(threading.enumerate())
+                            Autosubmit.check_logs_status(job_list, as_conf, new_run=False)
                             if job_list.get_failed():
                                 return 1
                             return 0
@@ -2281,6 +2266,7 @@ def run_experiment(expid, notransitive=False, start_time=None, start_after=None,
                                     "Couldn't recover the Historical database, AS will continue without it, GUI may be affected")
                         job_changes_tracker = {}
                         if Autosubmit.exit:
+                            Autosubmit.check_logs_status(job_list, as_conf, new_run=False)
                             job_list.save()
                             as_conf.save()
                         time.sleep(safetysleeptime)
@@ -2368,8 +2354,6 @@ def run_experiment(expid, notransitive=False, start_time=None, start_after=None,
                         raise # If this happens, there is a bug in the code or an exception not-well caught
                 Log.result("No more jobs to run.")
                 # search hint - finished run
-                for job in job_list.get_completed_failed_without_logs():
-                    job_list.update_log_status(job, as_conf)
                 job_list.save()
                 if not did_run and len(job_list.get_completed_failed_without_logs()) > 0: # Revise if there is any log unrecovered from previous run
                     Log.info(f"Connecting to the platforms, to recover missing logs")
@@ -2382,10 +2366,9 @@ def run_experiment(expid, notransitive=False, start_time=None, start_after=None,
                 Log.info("Waiting for all logs to be updated")
                 for p in platforms_to_test:
                     if p.log_recovery_process:
-                        p.cleanup_event.set() # Send cleanup event
+                        p.cleanup_event.set()  # Send cleanup event
                         p.log_recovery_process.join()
-                for job in job_list.get_completed_failed_without_logs(): # update the info gathered from the childs
-                    job_list.update_log_status(job, as_conf)
+                Autosubmit.check_logs_status(job_list, as_conf, new_run=False)
                 job_list.save()
                 if len(job_list.get_completed_failed_without_logs()) == 0:
                     Log.result(f"Autosubmit recovered all job logs.")
@@ -2401,8 +2384,8 @@ def run_experiment(expid, notransitive=False, start_time=None, start_after=None,
                         Autosubmit.database_fix(expid)
                     except Exception as e:
                         pass
-                for platform in platforms_to_test:
-                    platform.closeConnection()
+                for p in platforms_to_test:
+                    p.closeConnection()
                 if len(job_list.get_failed()) > 0:
                     Log.info("Some jobs have failed and reached maximum retrials")
                 else:
@@ -2573,7 +2556,6 @@ def submit_ready_jobs(as_conf, job_list, platforms_to_test, packages_persistence
             raise
         except BaseException as e:
             raise
-            raise AutosubmitCritical("This seems like a bug in the code, please contact AS developers", 7070, str(e))
 
     @staticmethod
     def monitor(expid, file_format, lst, filter_chunks, filter_status, filter_section, hide, txt_only=False,
@@ -2981,9 +2963,8 @@ def recovery(expid, noplot, save, all_jobs, hide, group_by=None, expand=list(),
                         "Experiment can't be recovered due being {0} active jobs in your experiment, If you want to recover the experiment, please use the flag -f and all active jobs will be cancelled".format(
                             len(current_active_jobs)), 7000)
             Log.debug("Job list restored from {0} files", pkl_dir)
-        except BaseException as e:
-            raise AutosubmitCritical(
-                "Couldn't restore the job_list or packages, check if the filesystem is having issues", 7040, str(e))
+        except Exception:
+            raise
         Log.info('Recovering experiment {0}'.format(expid))
         try:
             for job in job_list.get_job_list():
@@ -3017,13 +2998,8 @@ def recovery(expid, noplot, save, all_jobs, hide, group_by=None, expand=list(),
                     job.status = Status.COMPLETED
                     Log.info(
                         "CHANGED job '{0}' status to COMPLETED".format(job.name))
-                    # Log.status("CHANGED job '{0}' status to COMPLETED".format(job.name))
-
-                    if not no_recover_logs:
-                        try:
-                            job.platform.get_logs_files(expid, job.remote_logs)
-                        except Exception as e:
-                            pass
+                    job.recover_last_ready_date()
+                    job.recover_last_log_name()
                 elif job.status != Status.SUSPENDED:
                     job.status = Status.WAITING
                     job._fail_count = 0

autosubmit/job/job.py

@@ -222,6 +222,7 @@ def __init__(self, name, job_id, status, priority):
         self.parameters = None
         self._tmp_path = os.path.join(
             BasicConfig.LOCAL_ROOT_DIR, self.expid, BasicConfig.LOCAL_TMP_DIR)
+        self._log_path = Path(f"{self._tmp_path}/LOG_{self.expid}")
         self.write_start = False
         self._platform = None
         self.check = 'true'
@@ -274,7 +275,6 @@ def _init_runtime_parameters(self):
         self._memory_per_task = ''
         self.log_retrieved = False
         self.start_time_timestamp = time.time()
-        self.end_time_placeholder = time.time()
         self.processors_per_node = ""
         self.stat_file = self.script_name[:-4] + "_STAT_0"
 
@@ -2501,6 +2501,51 @@ def synchronize_logs(self, platform, remote_logs, local_logs, last = True):
             self.local_logs = local_logs
             self.remote_logs = copy.deepcopy(local_logs)
 
+    def _recover_last_log_name_from_filesystem(self) -> bool:
+        """
+        Recovers the log name for the job from the filesystem.
+        :return: True if the log name was already recovered, False otherwise
+        :rtype: bool
+        """
+        log_name = sorted(list(self._log_path.glob(f"{self.name}*")), key=lambda x: x.stat().st_mtime)
+        log_name = log_name[-1] if log_name else None
+        if log_name:
+            file_timestamp = int(datetime.datetime.fromtimestamp(log_name.stat().st_mtime).strftime("%Y%m%d%H%M%S"))
+            if self.ready_date and file_timestamp >= int(self.ready_date):
+                self.local_logs = (log_name.with_suffix(".out").name, log_name.with_suffix(".err").name)
+                self.remote_logs = copy.deepcopy(self.local_logs)
+                return True
+        self.local_logs = (f"{self.name}.out.{self._fail_count}", f"{self.name}.err.{self._fail_count}")
+        self.remote_logs = copy.deepcopy(self.local_logs)
+        return False
+
+    def recover_last_log_name(self):
+        """
+        Recovers the last log name for the job
+        """
+        if not self.updated_log:
+            self.updated_log = self._recover_last_log_name_from_filesystem()
+            # TODO: After PostgreSQL migration, implement _recover_last_log_from_db() to retrieve the last log from the database.
+
+    def recover_last_ready_date(self) -> None:
+        """
+        Recovers the last ready date for this job
+        """
+        if not self.ready_date:
+            stat_file = Path(f"{self._tmp_path}/{self.name}_TOTAL_STATS")
+            if stat_file.exists():
+                output_by_lines = stat_file.read_text().splitlines()
+                if output_by_lines:
+                    line_info = output_by_lines[-1].split(" ")
+                    if line_info and line_info[0].isdigit():
+                        self.ready_date = line_info[0]
+                    else:
+                        self.ready_date = datetime.datetime.fromtimestamp(stat_file.stat().st_mtime).strftime('%Y%m%d%H%M%S')
+                        Log.debug(f"Failed to recover ready date for the job {self.name}")
+                else:  # Default to last mod time
+                    self.ready_date = datetime.datetime.fromtimestamp(stat_file.stat().st_mtime).strftime('%Y%m%d%H%M%S')
+                    Log.debug(f"Failed to recover ready date for the job {self.name}")
+
 
 class WrapperJob(Job):
     """
@@ -2737,8 +2782,7 @@ def _check_running_jobs(self):
                 self._platform.send_file(multiple_checker_inner_jobs, False)
                 command = f"cd {self._platform.get_files_path()}; {os.path.join(self._platform.get_files_path(), 'inner_jobs_checker.sh')}"
             else:
-                command = os.path.join(
-                    self._platform.get_files_path(), "inner_jobs_checker.sh")
+                command = f"cd {self._platform.get_files_path()}; ./inner_jobs_checker.sh; cd {os.getcwd()}"
             #
             wait = 2
             retries = 5

autosubmit/job/job_list.py

@@ -2666,7 +2666,7 @@ def _count_parents_status(job: Job, target_status: str) -> Tuple[List[Job], List
                     non_completed_parents_current.append(parent[0])
         return non_completed_parents_current, completed_parents
 
-    def update_log_status(self, job, as_conf):
+    def update_log_status(self, job, as_conf, new_run=False):
         """
         Updates the log err and log out.
         """
@@ -2681,7 +2681,7 @@ def update_log_status(self, job, as_conf):
         if log_recovered:
             job.local_logs = (log_recovered.name, log_recovered.name[:-4] + ".err") # we only want the last one
             job.updated_log = True
-        elif not job.updated_log and str(as_conf.platforms_data.get(job.platform.name, {}).get('DISABLE_RECOVERY_THREADS', "false")).lower() == "false":
+        elif new_run and not job.updated_log and str(as_conf.platforms_data.get(job.platform.name, {}).get('DISABLE_RECOVERY_THREADS', "false")).lower() == "false":
             job.platform.add_job_to_log_recover(job)
         return log_recovered
 
@@ -2726,9 +2726,6 @@ def update_list(self, as_conf, store_change=True, fromSetStatus=False, submitter
         if self.update_from_file(store_change):
             save = store_change
         Log.debug('Updating FAILED jobs')
-        write_log_status = False
-        for job in self.get_completed_failed_without_logs():
-            save = self.update_log_status(job, as_conf) if not save else save
         if not first_time:
             for job in self.get_failed():
                 job.packed = False

autosubmit/job/job_packages.py

@@ -219,7 +219,13 @@ def _send_files(self):
     def _do_submission(self,job_scripts=None, hold=False):
         """ Submit package to the platform. """
 
-
+    def process_jobs_to_submit(self, job_id: str, hold: bool = False) -> None:
+        for i, job in enumerate(self.jobs):
+            job.hold = hold
+            job.id = str(job_id)
+            job.status = Status.SUBMITTED
+            if hasattr(self, "name"): # TODO change this check for a property that checks if it is a wrapper or not, the same change has to be done in other parts of the code
+                job.wrapper_name = self.name
 
 class JobPackageSimple(JobPackageBase):
     """
@@ -230,6 +236,7 @@ def __init__(self, jobs):
         super(JobPackageSimple, self).__init__(jobs)
         self._job_scripts = {}
         self.export = jobs[0].export
+        # self.name = "simple_package" TODO this should be possible, but it crashes accross the code. Add a property that defines what is a package with wrappers
 
     def _create_scripts(self, configuration):
         for job in self.jobs:

autosubmit/platforms/locplatform.py

@@ -163,7 +163,7 @@ def send_file(self, filename: str, check: bool = True) -> bool:
         Returns:
             bool: True if the file was sent successfully.
         """
-        command = f'{self.put_cmd} {os.path.join(self.tmp_path, Path(filename).name)} {os.path.join(self.tmp_path, "LOG_" + self.expid, Path(filename).name)}'
+        command = f'{self.put_cmd} {os.path.join(self.tmp_path, Path(filename).name)} {os.path.join(self.tmp_path, "LOG_" + self.expid, Path(filename).name)}; chmod 770 {os.path.join(self.tmp_path, "LOG_" + self.expid, Path(filename).name)}'
         try:
             subprocess.check_call(command, shell=True)
         except subprocess.CalledProcessError: