More fixes

Author: GitTimeraider

app.py

@@ -198,15 +198,24 @@ def schedule_all_repositories():
     except Exception as e:
         logger.error(f"Error scheduling repositories on startup: {e}")
 
-# Flag to ensure we only initialize once
+# Thread-safe flag to ensure we only initialize once
+import threading
+_scheduler_lock = threading.Lock()
 _scheduler_initialized = False
 
 def ensure_scheduler_initialized():
-    """Ensure scheduler is initialized with existing repositories"""
+    """Ensure scheduler is initialized with existing repositories (thread-safe)"""
     global _scheduler_initialized
-    if not _scheduler_initialized:
-        schedule_all_repositories()
-        _scheduler_initialized = True
+    if _scheduler_initialized:
+        return
+        
+    with _scheduler_lock:
+        # Double-check pattern to avoid race conditions
+        if not _scheduler_initialized:
+            logger.info("Initializing scheduler with existing repositories...")
+            schedule_all_repositories()
+            _scheduler_initialized = True
+            logger.info("Scheduler initialization completed")
 
 @login_manager.user_loader
 def load_user(user_id):
@@ -215,7 +224,6 @@ def load_user(user_id):
 @app.route('/')
 @login_required
 def dashboard():
-    ensure_scheduler_initialized()
     repositories = Repository.query.filter_by(user_id=current_user.id).all()
     recent_jobs = BackupJob.query.filter_by(user_id=current_user.id).order_by(BackupJob.created_at.desc()).limit(10).all()
     return render_template('dashboard.html', repositories=repositories, recent_jobs=recent_jobs)
@@ -647,14 +655,15 @@ def backup_with_context():
                 ).first()
 
                 if running_job:
-                    logger.warning(f"Backup already running for repository {repo.name}, skipping")
+                    logger.warning(f"Backup already running for repository {repo.name} (job {running_job.id}), skipping")
                     return
 
-                # Additional check: ensure no backup started in the last 5 minutes to prevent rapid duplicates
+                # Additional check: ensure no backup started in the last 30 seconds to prevent rapid duplicates
+                recent_cutoff = datetime.utcnow() - timedelta(seconds=30)
                 recent_backup = BackupJob.query.filter_by(
                     repository_id=repository.id
                 ).filter(
-                    BackupJob.started_at > datetime.utcnow() - timedelta(minutes=5)
+                    BackupJob.started_at > recent_cutoff
                 ).first()
 
                 if recent_backup:
@@ -754,5 +763,15 @@ def backup_with_context():
     else:
         logger.error(f"Failed to schedule job {job_id}")
 
+# Initialize scheduler with existing repositories at startup
+# This runs after all functions are defined
+try:
+    with app.app_context():
+        logger.info("Starting scheduler initialization at app startup...")
+        ensure_scheduler_initialized()
+        logger.info("Scheduler initialization at startup completed")
+except Exception as e:
+    logger.error(f"Failed to initialize scheduler at startup: {e}")
+
 if __name__ == '__main__':
     app.run(host='0.0.0.0', port=8080, debug=False)

backup_service.py

@@ -4,7 +4,7 @@
 import zipfile
 import tarfile
 import logging
-from datetime import datetime
+from datetime import datetime, timedelta
 from pathlib import Path
 from github import Github
 from models import db, BackupJob
@@ -21,6 +21,28 @@ def backup_repository(self, repository):
         """Backup a repository according to its settings"""
         logger.info(f"Starting backup for repository: {repository.name}")
 
+        # Check if there's already a running backup for this repository
+        existing_running_job = BackupJob.query.filter_by(
+            repository_id=repository.id,
+            status='running'
+        ).first()
+        
+        if existing_running_job:
+            logger.warning(f"Backup already running for repository {repository.name} (job {existing_running_job.id}), skipping")
+            return
+        
+        # Also check for very recent backups (within last 30 seconds) to prevent rapid duplicates
+        recent_cutoff = datetime.utcnow() - timedelta(seconds=30)
+        recent_job = BackupJob.query.filter_by(
+            repository_id=repository.id
+        ).filter(
+            BackupJob.started_at > recent_cutoff
+        ).first()
+        
+        if recent_job:
+            logger.warning(f"Very recent backup found for repository {repository.name} (started at {recent_job.started_at}), skipping to prevent duplicates")
+            return
+        
         # Create backup job record
         backup_job = BackupJob(
             user_id=repository.user_id,
@@ -29,7 +51,15 @@ def backup_repository(self, repository):
             started_at=datetime.utcnow()
         )
         db.session.add(backup_job)
-        db.session.commit()
+        
+        # Commit immediately to make this job visible to other processes/threads
+        try:
+            db.session.commit()
+            logger.info(f"Created backup job {backup_job.id} for repository {repository.name}")
+        except Exception as e:
+            logger.error(f"Failed to commit backup job creation: {e}")
+            db.session.rollback()
+            return
 
         temp_clone_dir = None
         try: