From 7a01ad76143973199bd6965c13476d2d04f10f75 Mon Sep 17 00:00:00 2001
From: Blake Mizerany <blake.mizerany@gmail.com>
Date: Mon, 3 Mar 2025 19:11:16 -0800
Subject: [PATCH] server/internal/registry: reintroduce pruning on model
 deletion (#9489)

This reintroduces aggressive pruning on model deletion as a temporary
measure until a more controlled garbage collection (GC) mechanism is
implemented.

Issues with the current approach:

1. Users may accidentally delete a model (`ollama rm llama3.3` instead
   of `ollama rm llama3.2`), requiring a full re-download unless another
   model references the same blobs.

2. Users may assume a deleted model is still referenced elsewhere, but
   due to prior updates or deletions, the references no longer exist,
   leading to unnecessary re-downloads.

Soon, we should implement a structured GC mechanism to retain
unreferenced blobs for a configurable period before removal, which will
run on "ollama rm" and other commands we deem appropriate.

Users that want to immediately remove unreferenced blobs can use a new
prune command that will allow them to specify the age and class of blobs
to remove.

Example usage:

    # Run basic blob GC
    $ ollama prune

    # Remove unreferenced blobs older than 7 days
    $ ollama prune --age 7d

    # Remove all blobs, referenced or not, older than 7 days (and their manifests?)
    $ ollama prune --age 7d --all

    # Remove all unreferenced blobs immediately
    $ ollama prune --age 0 --all

    # Remove all blobs
    $ ollama prune --age 0 --all

This should provide a safer and more predictable cleanup process.
---
 server/internal/registry/server.go | 9 ++++++++-
 server/routes.go                   | 2 ++
 2 files changed, 10 insertions(+), 1 deletion(-)

diff --git a/server/internal/registry/server.go b/server/internal/registry/server.go
index 81085357d..4d44aa8d0 100644
--- a/server/internal/registry/server.go
+++ b/server/internal/registry/server.go
@@ -31,6 +31,10 @@ type Local struct {
 	// Fallback, if set, is used to handle requests that are not handled by
 	// this handler.
 	Fallback http.Handler
+
+	// Prune, if set, is called to prune the local disk cache after a model
+	// is deleted.
+	Prune func() error // optional
 }
 
 // serverError is like ollama.Error, but with a Status field for the HTTP
@@ -204,7 +208,10 @@ func (s *Local) handleDelete(_ http.ResponseWriter, r *http.Request) error {
 	if !ok {
 		return &serverError{404, "not_found", "model not found"}
 	}
-	return nil
+	if s.Prune == nil {
+		return nil
+	}
+	return s.Prune()
 }
 
 func decodeUserJSON[T any](r io.Reader) (T, error) {
diff --git a/server/routes.go b/server/routes.go
index 519f04738..f1afad6e0 100644
--- a/server/routes.go
+++ b/server/routes.go
@@ -1199,6 +1199,8 @@ func (s *Server) GenerateRoutes(rc *ollama.Registry) (http.Handler, error) {
 		Client:   rc,
 		Logger:   slog.Default(), // TODO(bmizerany): Take a logger, do not use slog.Default()
 		Fallback: r,
+
+		Prune: PruneLayers,
 	}
 
 	return rs, nil