Pleroma.Upload.Filter.Dedupe: sharding directory structure

Dedupe now uses a three-level sharding directory structure to improve performance when many files are uploaded and stored on a filesystem instead of an object store. (note: Minio still affected as it still uses a traditional filesystem)

This does not help if you already have hundreds of thousands of files uploaded. The media URLs are permanently part of the activity so the files cannot be relocated. A motivated user could write a tool to move the files and perhaps write an Nginx or equivalent redirect to make the files still accessible, but that is beyond the scope of this change.

5 files changed, 21 insertions, 8 deletions

diff --git a/changelog.d/dedupe-sharding.change b/changelog.d/dedupe-sharding.change
new file mode 100644
index 000000000..2e140d8a2
--- /dev/null
+++ b/changelog.d/dedupe-sharding.change
@@ -0,0 +1 @@
+Dedupe upload filter now uses a three-level sharding directory structure
diff --git a/lib/pleroma/upload/filter/dedupe.ex b/lib/pleroma/upload/filter/dedupe.ex
index ef793d390..7b278d299 100644
--- a/lib/pleroma/upload/filter/dedupe.ex
+++ b/lib/pleroma/upload/filter/dedupe.ex
@@ -17,8 +17,16 @@ defmodule Pleroma.Upload.Filter.Dedupe do
       |> Base.encode16(case: :lower)
 
     filename = shasum <> "." <> extension
-    {:ok, :filtered, %Upload{upload | id: shasum, path: filename}}
+
+    {:ok, :filtered, %Upload{upload | id: shasum, path: shard_path(filename)}}
   end
 
   def filter(_), do: {:ok, :noop}
+
+  @spec shard_path(String.t()) :: String.t()
+  def shard_path(
+        <<a::binary-size(2), b::binary-size(2), c::binary-size(2), _::binary>> = filename
+      ) do
+    Path.join([a, b, c, filename])
+  end
 end
diff --git a/test/pleroma/object_test.exs b/test/pleroma/object_test.exs
index b3c528e32..ed5c2b6c8 100644
--- a/test/pleroma/object_test.exs
+++ b/test/pleroma/object_test.exs
@@ -174,8 +174,9 @@ defmodule Pleroma.ObjectTest do
 
       filename = Path.basename(href)
 
-      assert {:ok, files} = File.ls(uploads_dir)
-      assert filename in files
+      expected_path = Path.join([uploads_dir, Pleroma.Upload.Filter.Dedupe.shard_path(filename)])
+
+      assert File.exists?(expected_path)
 
       Object.delete(note)
 
@@ -183,8 +184,7 @@ defmodule Pleroma.ObjectTest do
 
       assert Object.get_by_id(note.id).data["deleted"]
       assert Object.get_by_id(attachment.id) == nil
-      assert {:ok, files} = File.ls(uploads_dir)
-      refute filename in files
+      refute File.exists?(expected_path)
     end
 
     test "with objects that have legacy data.url attribute" do
diff --git a/test/pleroma/upload/filter/dedupe_test.exs b/test/pleroma/upload/filter/dedupe_test.exs
index 29c181509..cd5ce121b 100644
--- a/test/pleroma/upload/filter/dedupe_test.exs
+++ b/test/pleroma/upload/filter/dedupe_test.exs
@@ -23,10 +23,12 @@ defmodule Pleroma.Upload.Filter.DedupeTest do
       tempfile: Path.absname("test/fixtures/image_tmp.jpg")
     }
 
+    expected_path = Dedupe.shard_path(@shasum <> ".jpg")
+
     assert {
              :ok,
              :filtered,
-             %Pleroma.Upload{id: @shasum, path: @shasum <> ".jpg"}
+             %Pleroma.Upload{id: @shasum, path: ^expected_path}
            } = Dedupe.filter(upload)
   end
 end
diff --git a/test/pleroma/upload_test.exs b/test/pleroma/upload_test.exs
index facb634c3..5fd62fa43 100644
--- a/test/pleroma/upload_test.exs
+++ b/test/pleroma/upload_test.exs
@@ -149,6 +149,9 @@ defmodule Pleroma.UploadTest do
 
     test "copies the file to the configured folder with deduping" do
       File.cp!("test/fixtures/image.jpg", "test/fixtures/image_tmp.jpg")
+      expected_filename = "e30397b58d226d6583ab5b8b3c5defb0c682bda5c31ef07a9f57c1c4986e3781.jpg"
+
+      expected_path = Pleroma.Upload.Filter.Dedupe.shard_path(expected_filename)
 
       file = %Plug.Upload{
         content_type: "image/jpeg",
@@ -159,8 +162,7 @@ defmodule Pleroma.UploadTest do
       {:ok, data} = Upload.store(file, filters: [Pleroma.Upload.Filter.Dedupe])
 
       assert List.first(data["url"])["href"] ==
-               Pleroma.Upload.base_url() <>
-                 "e30397b58d226d6583ab5b8b3c5defb0c682bda5c31ef07a9f57c1c4986e3781.jpg"
+               Path.join([Pleroma.Upload.base_url(), expected_path])
     end
 
     test "copies the file to the configured folder without deduping" do