diff --git a/swh/objstorage/backends/libcloud.py b/swh/objstorage/backends/libcloud.py
--- a/swh/objstorage/backends/libcloud.py
+++ b/swh/objstorage/backends/libcloud.py
@@ -142,7 +142,12 @@
 
     def get(self, obj_id):
         obj = b''.join(self._get_object(obj_id).as_stream())
-        return decompressors[self.compression](obj)
+        d = decompressors[self.compression]()
+        ret = d.decompress(obj)
+        if d.unused_data:
+            hex_obj_id = hashutil.hash_to_hex(obj_id)
+            raise Error('Corrupt object %s: trailing data found' % hex_obj_id)
+        return ret
 
     def check(self, obj_id):
         # Check that the file exists, as _get_object raises ObjNotFoundError
diff --git a/swh/objstorage/backends/seaweed.py b/swh/objstorage/backends/seaweed.py
--- a/swh/objstorage/backends/seaweed.py
+++ b/swh/objstorage/backends/seaweed.py
@@ -149,10 +149,16 @@
     def get(self, obj_id):
         try:
             obj = self.wf.get(self._path(obj_id))
-            return decompressors[self.compression](obj)
         except Exception:
             raise ObjNotFoundError(obj_id)
 
+        d = decompressors[self.compression]()
+        ret = d.decompress(obj)
+        if d.unused_data:
+            hex_obj_id = hashutil.hash_to_hex(obj_id)
+            raise Error('Corrupt object %s: trailing data found' % hex_obj_id)
+        return ret
+
     def check(self, obj_id):
         # Check the content integrity
         obj_content = self.get(obj_id)
diff --git a/swh/objstorage/objstorage.py b/swh/objstorage/objstorage.py
--- a/swh/objstorage/objstorage.py
+++ b/swh/objstorage/objstorage.py
@@ -6,7 +6,6 @@
 import abc
 from itertools import dropwhile, islice
 import bz2
-import gzip
 import lzma
 import zlib
 
@@ -46,13 +45,22 @@
         return b''
 
 
+class NullDecompressor:
+    def decompress(self, data):
+        return data
+
+    @property
+    def unused_data(self):
+        return b''
+
+
 decompressors = {
-    'bz2': bz2.decompress,
-    'lzma': lzma.decompress,
-    'gzip': gzip.decompress,
-    'zlib': zlib.decompress,
-    None: lambda x: x,
-    }
+    'bz2': bz2.BZ2Decompressor,
+    'lzma': lzma.LZMADecompressor,
+    'gzip': lambda: zlib.decompressobj(wbits=31),
+    'zlib': zlib.decompressobj,
+    None: NullDecompressor,
+}
 
 compressors = {
     'bz2': bz2.BZ2Compressor,
diff --git a/swh/objstorage/tests/test_objstorage_cloud.py b/swh/objstorage/tests/test_objstorage_cloud.py
--- a/swh/objstorage/tests/test_objstorage_cloud.py
+++ b/swh/objstorage/tests/test_objstorage_cloud.py
@@ -111,7 +111,9 @@
         obj_id = hashutil.hash_to_hex(obj_id)
         raw_content = b''.join(data[obj_id].content)
 
-        assert decompressors[self.compression](raw_content) == content
+        d = decompressors[self.compression]()
+        assert d.decompress(raw_content) == content
+        assert d.unused_data == b''
 
 
 class TestCloudObjStorageBz2(TestCloudObjStorage):