check: report missing chunks grouped as chunk -> files -> archives, #9218

src/borg/archive.py

@@ -2075,6 +2075,9 @@ def rebuild_archives(
     ):
         """Analyze and rebuild archives, expecting some damage and trying to make stuff consistent again."""
 
+        missing_chunk_size: dict = {}  # chunk_id -> chunk size in bytes
+        missing_chunk_refs: defaultdict = defaultdict(lambda: defaultdict(set))  # chunk_id -> {path: {archive_name}}
+
         def add_callback(chunk):
             id_ = self.key.id_hash(chunk)
             cdata = self.repo_objs.format(id_, {}, chunk, ro_type=ROBJ_ARCHIVE_STREAM)
@@ -2091,16 +2094,22 @@ def add_reference(id_, size, cdata):
                     self.chunks.update_pack_info(pack_results)
 
         def verify_file_chunks(archive_name, item):
-            """Verifies that all file chunks are present. Missing file chunks will be logged."""
+            """Verify that all file chunks are present.
+
+            Record each missing chunk's size in missing_chunk_size and, in missing_chunk_refs, the
+            file path and archive it occurs in. Log each missing chunk at debug level.
+            """
             offset = 0
             for chunk in item.chunks:
                 chunk_id, size = chunk
                 if chunk_id not in self.chunks:
-                    logger.error(
+                    logger.debug(
                         "{}: {}: Missing file chunk detected (Byte {}-{}, Chunk {}).".format(
                             archive_name, item.path, offset, offset + size, bin_to_hex(chunk_id)
                         )
                     )
+                    missing_chunk_size[chunk_id] = size
+                    missing_chunk_refs[chunk_id][item.path].add(archive_name)
                     self.error_found = True
                 offset += size
             if "size" in item:
@@ -2114,6 +2123,17 @@ def verify_file_chunks(archive_name, item):
                         )
                     )
 
+        def report_missing_chunks():
+            """Log the missing chunks, each with its size and the files and archives referencing it."""
+            if not missing_chunk_refs:
+                return
+            logger.error("The following chunks are missing in the repository:")
+            for chunk_id, refs in missing_chunk_refs.items():
+                logger.error(f"- Chunk {bin_to_hex(chunk_id)}, {missing_chunk_size[chunk_id]:,} bytes")
+                for path in sorted(refs):
+                    archive_names = ", ".join(sorted(refs[path]))
+                    logger.error(f"    - {path}: {archive_names}")
+
         def robust_iterator(archive):
             """Iterates through all archive items
 
@@ -2271,6 +2291,7 @@ def valid_item(obj):
                 if archive_id != new_archive_id:
                     self.manifest.archives.delete_by_id(archive_id)
         pi.finish()
+        report_missing_chunks()
 
     def finish(self):
         if self.repair:

src/borg/testsuite/archiver/check_cmd_test.py

@@ -169,9 +169,13 @@ def test_missing_file_chunk(archivers, request):
             pytest.fail("should not happen")  # convert 'fail'
 
     output = cmd(archiver, "check", exit_code=1)
-    assert "Missing file chunk detected" in output
+    assert "The following chunks are missing in the repository:" in output
+    assert bin_to_hex(killed_chunk.id) in output
+    assert src_file in output
     output = cmd(archiver, "check", "--repair", exit_code=0)
-    assert "Missing file chunk detected" in output  # repair is not changing anything, just reporting.
+    # repair is not changing anything, just reporting.
+    assert "The following chunks are missing in the repository:" in output
+    assert bin_to_hex(killed_chunk.id) in output
 
     # check does not modify the chunks list.
     for archive_name in ("archive1", "archive2"):
@@ -191,7 +195,7 @@ def test_missing_file_chunk(archivers, request):
 
     # check should not complain anymore about missing chunks:
     output = cmd(archiver, "check", "-v", "--repair", exit_code=0)
-    assert "Missing file chunk detected" not in output
+    assert "The following chunks are missing in the repository:" not in output
 
 
 def test_missing_archive_item_chunk(archivers, request):
@@ -474,11 +478,14 @@ def test_verify_data(archivers, request, init_args):
     # repair will find the defect chunk and remove it
     output = cmd(archiver, "check", "--repair", "--verify-data", exit_code=0)
     assert f"{bin_to_hex(chunk.id)}, integrity error" in output
-    assert f"{src_file}: Missing file chunk detected" in output
+    assert "The following chunks are missing in the repository:" in output
+    assert bin_to_hex(chunk.id) in output
+    assert src_file in output
 
     # run with --verify-data again, it will notice the missing chunk.
     output = cmd(archiver, "check", "--archives-only", "--verify-data", exit_code=1)
-    assert f"{src_file}: Missing file chunk detected" in output
+    assert "The following chunks are missing in the repository:" in output
+    assert bin_to_hex(chunk.id) in output
 
 
 @pytest.mark.parametrize("init_args", [["--encryption=aes256-ocb"], ["--encryption", "none"]])
@@ -506,12 +513,15 @@ def test_corrupted_file_chunk(archivers, request, init_args):
     # repair: the defect chunk will be removed.
     output = cmd(archiver, "check", "--repair", "--verify-data", exit_code=0)
     assert f"{bin_to_hex(chunk.id)}, integrity error" in output
-    assert f"{src_file}: Missing file chunk detected" in output
+    assert "The following chunks are missing in the repository:" in output
+    assert bin_to_hex(chunk.id) in output
+    assert src_file in output
 
     # run normal check again
     cmd(archiver, "check", "--repository-only", exit_code=0)
     output = cmd(archiver, "check", "--archives-only", exit_code=1)
-    assert f"{src_file}: Missing file chunk detected" in output
+    assert "The following chunks are missing in the repository:" in output
+    assert src_file in output
 
 
 @pytest.mark.skip(