Recover disk space safely and complete metadata-only backup updates

lib/space.py

@@ -9,6 +9,7 @@ import os
 from pathlib import Path
 import re
 import shutil
+import subprocess
 
 import backup_set
 
@@ -18,12 +19,21 @@ def private(path):
             path.stat().st_uid == 0 and not path.stat().st_mode & 0o077)
 
 
+def mounted_paths():
+    # st_dev/ismount alone miss same-filesystem bind mounts. Linux mountinfo
+    # lists file and directory mounts, escaping whitespace and backslashes.
+    return {Path(re.sub(r'\\([0-7]{3})', lambda m: chr(int(m[1], 8)), line.split()[4]))
+            for line in Path('/proc/self/mountinfo').read_text().splitlines()}
+
+
 def remove_tree(path):
     if not private(path):
         raise ValueError('Refusing non-private/aliased scratch: ' + str(path))
+    if os.path.ismount(path) or any(p == path or path in p.parents for p in mounted_paths()):
+        raise ValueError('Refusing mounted scratch: ' + str(path))
     device = path.stat().st_dev
-    for parent, dirs, _ in os.walk(path, followlinks=False):
-        for name in dirs:
+    for parent, dirs, files in os.walk(path, followlinks=False):
+        for name in dirs + files:
             child = Path(parent) / name
             if not child.is_symlink() and (child.stat().st_dev != device or os.path.ismount(child)):
                 raise ValueError('Refusing mounted scratch: ' + str(child))
@@ -66,7 +76,9 @@ def clean(work, staging, archives, prefix, restore=None):
                 print('-- Removed abandoned verification download: ' + path.name, flush=True)
 
 
-def eligible(archives, prefix, keep):
+def uploaded_sets(archives, prefix):
+    if not private(archives):
+        raise ValueError('Cleanup requires private canonical archives: ' + str(archives))
     valid = []
     for path in archives.iterdir():
         if not private(path) or not path.name.startswith(prefix+'_'):
@@ -81,19 +93,48 @@ def eligible(archives, prefix, keep):
         except (ValueError, OSError, KeyError, TypeError):
             continue
         valid.append(path.name)
-    # Emergency pruning may ignore age, NEVER the minimum newest-copy count.
-    for name in sorted(valid, reverse=True)[keep:][::-1]:
+    return sorted(valid, reverse=True)
+
+
+def eligible(archives, prefix, keep, protected=False):
+    if keep < 1:
+        raise ValueError('At least one uploaded local set must be retained')
+    # Pressure policy is independent of normal age/count retention.
+    valid = uploaded_sets(archives, prefix)
+    for name in (valid[:keep] if protected else valid[keep:][::-1]):
         print(name)
 
 
+def inventory(archives, prefix, keep):
+    valid = uploaded_sets(archives, prefix)
+    protected = set(valid[:keep])
+    print('-- Remaining local archive inventory (allocated bytes; no deletion):', flush=True)
+    for path in sorted(archives.iterdir()):
+        if not private(path):
+            print(f'   {path.name}: protected unknown/aliased/non-private entry', flush=True)
+            continue
+        if path.name in protected:
+            reason = 'protected local minimum'
+        elif path.name in valid:
+            reason = 'uploaded; retained (verification/source/device constraints)'
+        else:
+            reason = 'protected unconfirmed/incomplete/unknown set'
+        result = subprocess.run(['du', '-sx', '-B1', '--', str(path)],
+                                text=True, capture_output=True, check=True)
+        size = int(result.stdout.split()[0])
+        print(f'   {path.name}: {size} bytes; {reason}', flush=True)
+
+
 if __name__ == '__main__':
     parser = argparse.ArgumentParser()
-    parser.add_argument('mode', choices=('clean', 'eligible', 'remove'))
+    parser.add_argument('mode', choices=('clean', 'eligible', 'protected', 'inventory', 'remove'))
     parser.add_argument('paths', nargs='+')
     args = parser.parse_args()
     if args.mode == 'clean':
         clean(*(Path(p) for p in args.paths[:3]), args.paths[3], Path(args.paths[4]) if len(args.paths) > 4 else None)
-    elif args.mode == 'eligible':
-        eligible(Path(args.paths[0]), args.paths[1], int(args.paths[2]))
+    elif args.mode in ('eligible', 'protected'):
+        eligible(Path(args.paths[0]), args.paths[1], int(args.paths[2]), args.mode == 'protected')
+    elif args.mode == 'inventory':
+        inventory(Path(args.paths[0]), args.paths[1], int(args.paths[2]))
     else:
         remove_tree(Path(args.paths[0]))