[feature] Cleanup unattached local media (#680)

* add localUnattached db function

* add parseOlderThan util function

* add pruneunusedlocalattachments to media manager

* add unusedlocal pruning to schedule + admin call

* set number of days to keep as a const

* fix test

internal/db/bundb/media.go

@@ -98,3 +98,29 @@ func (m *mediaDB) GetAvatarsAndHeaders(ctx context.Context, maxID string, limit
 
 	return attachments, nil
 }
+
+func (m *mediaDB) GetLocalUnattachedOlderThan(ctx context.Context, olderThan time.Time, maxID string, limit int) ([]*gtsmodel.MediaAttachment, db.Error) {
+	attachments := []*gtsmodel.MediaAttachment{}
+
+	q := m.newMediaQ(&attachments).
+		Where("media_attachment.cached = true").
+		Where("media_attachment.avatar = false").
+		Where("media_attachment.header = false").
+		Where("media_attachment.created_at < ?", olderThan).
+		Where("media_attachment.remote_url IS NULL").
+		Where("media_attachment.status_id IS NULL")
+
+	if maxID != "" {
+		q = q.Where("media_attachment.id < ?", maxID)
+	}
+
+	if limit != 0 {
+		q = q.Limit(limit)
+	}
+
+	if err := q.Scan(ctx); err != nil {
+		return nil, m.conn.ProcessError(err)
+	}
+
+	return attachments, nil
+}

internal/db/bundb/media_test.go

@@ -24,6 +24,7 @@ import (
 	"time"
 
 	"github.com/stretchr/testify/suite"
+	"github.com/superseriousbusiness/gotosocial/testrig"
 )
 
 type MediaTestSuite struct {
@@ -51,6 +52,14 @@ func (suite *MediaTestSuite) TestGetAvisAndHeaders() {
 	suite.Len(attachments, 2)
 }
 
+func (suite *MediaTestSuite) TestGetLocalUnattachedOlderThan() {
+	ctx := context.Background()
+
+	attachments, err := suite.db.GetLocalUnattachedOlderThan(ctx, testrig.TimeMustParse("2090-06-04T13:12:00Z"), "", 10)
+	suite.NoError(err)
+	suite.Len(attachments, 1)
+}
+
 func TestMediaTestSuite(t *testing.T) {
 	suite.Run(t, new(MediaTestSuite))
 }

internal/db/media.go

@@ -38,4 +38,9 @@ type Media interface {
 	// GetAvatarsAndHeaders fetches limit n avatars and headers with an id < maxID. These headers
 	// and avis may be in use or not; the caller should check this if it's important.
 	GetAvatarsAndHeaders(ctx context.Context, maxID string, limit int) ([]*gtsmodel.MediaAttachment, Error)
+	// GetLocalUnattachedOlderThan fetches limit n local media attachments, older than the given time, which
+	// aren't header or avatars, and aren't attached to a status. In other words, attachments which were uploaded
+	// but never used for whatever reason, or attachments that were attached to a status which was subsequently
+	// deleted.
+	GetLocalUnattachedOlderThan(ctx context.Context, olderThan time.Time, maxID string, limit int) ([]*gtsmodel.MediaAttachment, Error)
 }

internal/media/manager.go

@@ -34,6 +34,10 @@ import (
 // selectPruneLimit is the amount of media entries to select at a time from the db when pruning
 const selectPruneLimit = 20
 
+// UnusedLocalAttachmentCacheDays is the amount of days to keep local media in storage if it
+// is not attached to a status, or was never attached to a status.
+const UnusedLocalAttachmentCacheDays = 3
+
 // Manager provides an interface for managing media: parsing, storing, and retrieving media objects like photos, videos, and gifs.
 type Manager interface {
 	// ProcessMedia begins the process of decoding and storing the given data as an attachment.
@@ -75,11 +79,16 @@ type Manager interface {
 	//
 	// The returned int is the amount of media that was pruned by this function.
 	PruneAllRemote(ctx context.Context, olderThanDays int) (int, error)
-	// PruneAllMeta prunes unused meta media -- currently, this means unused avatars + headers, but can also be extended
+	// PruneAllMeta prunes unused/out of date headers and avatars cached on this instance.
-	// to include things like attachments that were uploaded on this server but left unused, etc.
 	//
 	// The returned int is the amount of media that was pruned by this function.
 	PruneAllMeta(ctx context.Context) (int, error)
+	// PruneUnusedLocalAttachments prunes unused media attachments that were uploaded by
+	// a user on this instance, but never actually attached to a status, or attached but
+	// later detached.
+	//
+	// The returned int is the amount of media that was pruned by this function.
+	PruneUnusedLocalAttachments(ctx context.Context) (int, error)
 
 	// Stop stops the underlying worker pool of the manager. It should be called
 	// when closing GoToSocial in order to cleanly finish any in-progress jobs.
@@ -210,6 +219,19 @@ func scheduleCleanupJobs(m *manager) error {
 		return fmt.Errorf("error starting media manager meta cleanup job: %s", err)
 	}
 
+	if _, err := c.AddFunc("@midnight", func() {
+		begin := time.Now()
+		pruned, err := m.PruneUnusedLocalAttachments(pruneCtx)
+		if err != nil {
+			logrus.Errorf("media manager: error pruning unused local attachments: %s", err)
+			return
+		}
+		logrus.Infof("media manager: pruned %d unused local attachments in %s", pruned, time.Since(begin))
+	}); err != nil {
+		pruneCancel()
+		return fmt.Errorf("error starting media manager unused local attachments cleanup job: %s", err)
+	}
+
 	// start remote cache cleanup cronjob if configured
 	if mediaRemoteCacheDays := config.GetMediaRemoteCacheDays(); mediaRemoteCacheDays > 0 {
 		if _, err := c.AddFunc("@midnight", func() {

internal/media/pruneremote.go

@@ -21,7 +21,6 @@ package media
 import (
 	"context"
 	"fmt"
-	"time"
 
 	"codeberg.org/gruf/go-store/storage"
 	"github.com/sirupsen/logrus"
@@ -32,15 +31,10 @@ import (
 func (m *manager) PruneAllRemote(ctx context.Context, olderThanDays int) (int, error) {
 	var totalPruned int
 
-	// convert days into a duration string
+	olderThan, err := parseOlderThan(olderThanDays)
-	olderThanHoursString := fmt.Sprintf("%dh", olderThanDays*24)
-	// parse the duration string into a duration
-	olderThanHours, err := time.ParseDuration(olderThanHoursString)
 	if err != nil {
-		return totalPruned, fmt.Errorf("PruneAllRemote: %d", err)
+		return totalPruned, fmt.Errorf("PruneAllRemote: error parsing olderThanDays %d: %s", olderThanDays, err)
 	}
-	// 'subtract' that from the time now to give our threshold
-	olderThan := time.Now().Add(-olderThanHours)
 	logrus.Infof("PruneAllRemote: pruning media older than %s", olderThan)
 
 	// select 20 attachments at a time and prune them

internal/media/pruneunusedlocal.go

@@ -0,0 +1,86 @@
+/*
+   GoToSocial
+   Copyright (C) 2021-2022 GoToSocial Authors admin@gotosocial.org
+
+   This program is free software: you can redistribute it and/or modify
+   it under the terms of the GNU Affero General Public License as published by
+   the Free Software Foundation, either version 3 of the License, or
+   (at your option) any later version.
+
+   This program is distributed in the hope that it will be useful,
+   but WITHOUT ANY WARRANTY; without even the implied warranty of
+   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+   GNU Affero General Public License for more details.
+
+   You should have received a copy of the GNU Affero General Public License
+   along with this program.  If not, see <http://www.gnu.org/licenses/>.
+*/
+
+package media
+
+import (
+	"context"
+	"fmt"
+
+	"codeberg.org/gruf/go-store/storage"
+	"github.com/sirupsen/logrus"
+	"github.com/superseriousbusiness/gotosocial/internal/db"
+	"github.com/superseriousbusiness/gotosocial/internal/gtsmodel"
+)
+
+func (m *manager) PruneUnusedLocalAttachments(ctx context.Context) (int, error) {
+	var totalPruned int
+	var maxID string
+	var attachments []*gtsmodel.MediaAttachment
+	var err error
+
+	olderThan, err := parseOlderThan(UnusedLocalAttachmentCacheDays)
+	if err != nil {
+		return totalPruned, fmt.Errorf("PruneUnusedLocalAttachments: error parsing olderThanDays %d: %s", UnusedLocalAttachmentCacheDays, err)
+	}
+	logrus.Infof("PruneUnusedLocalAttachments: pruning unused local attachments older than %s", olderThan)
+
+	// select 20 attachments at a time and prune them
+	for attachments, err = m.db.GetLocalUnattachedOlderThan(ctx, olderThan, maxID, selectPruneLimit); err == nil && len(attachments) != 0; attachments, err = m.db.GetLocalUnattachedOlderThan(ctx, olderThan, maxID, selectPruneLimit) {
+		// use the id of the last attachment in the slice as the next 'maxID' value
+		l := len(attachments)
+		maxID = attachments[l-1].ID
+		logrus.Tracef("PruneUnusedLocalAttachments: got %d unused local attachments older than %s with maxID < %s", l, olderThan, maxID)
+
+		for _, attachment := range attachments {
+			if err := m.pruneOneLocal(ctx, attachment); err != nil {
+				return totalPruned, err
+			}
+			totalPruned++
+		}
+	}
+
+	// make sure we don't have a real error when we leave the loop
+	if err != nil && err != db.ErrNoEntries {
+		return totalPruned, err
+	}
+
+	logrus.Infof("PruneUnusedLocalAttachments: finished pruning: pruned %d entries", totalPruned)
+	return totalPruned, nil
+}
+
+func (m *manager) pruneOneLocal(ctx context.Context, attachment *gtsmodel.MediaAttachment) error {
+	if attachment.File.Path != "" {
+		// delete the full size attachment from storage
+		logrus.Tracef("pruneOneLocal: deleting %s", attachment.File.Path)
+		if err := m.storage.Delete(attachment.File.Path); err != nil && err != storage.ErrNotFound {
+			return err
+		}
+	}
+
+	if attachment.Thumbnail.Path != "" {
+		// delete the thumbnail from storage
+		logrus.Tracef("pruneOneLocal: deleting %s", attachment.Thumbnail.Path)
+		if err := m.storage.Delete(attachment.Thumbnail.Path); err != nil && err != storage.ErrNotFound {
+			return err
+		}
+	}
+
+	// delete the attachment completely
+	return m.db.DeleteByID(ctx, attachment.ID, attachment)
+}

internal/media/pruneunusedlocal_test.go

@@ -0,0 +1,75 @@
+/*
+   GoToSocial
+   Copyright (C) 2021-2022 GoToSocial Authors admin@gotosocial.org
+
+   This program is free software: you can redistribute it and/or modify
+   it under the terms of the GNU Affero General Public License as published by
+   the Free Software Foundation, either version 3 of the License, or
+   (at your option) any later version.
+
+   This program is distributed in the hope that it will be useful,
+   but WITHOUT ANY WARRANTY; without even the implied warranty of
+   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+   GNU Affero General Public License for more details.
+
+   You should have received a copy of the GNU Affero General Public License
+   along with this program.  If not, see <http://www.gnu.org/licenses/>.
+*/
+
+package media_test
+
+import (
+	"context"
+	"testing"
+
+	"github.com/stretchr/testify/suite"
+	"github.com/superseriousbusiness/gotosocial/internal/db"
+)
+
+type PruneUnusedLocalTestSuite struct {
+	MediaStandardTestSuite
+}
+
+func (suite *PruneUnusedLocalTestSuite) TestPruneUnusedLocal() {
+	testAttachment := suite.testAttachments["local_account_1_unattached_1"]
+	suite.True(testAttachment.Cached)
+
+	totalPruned, err := suite.manager.PruneUnusedLocalAttachments(context.Background())
+	suite.NoError(err)
+	suite.Equal(1, totalPruned)
+
+	_, err = suite.db.GetAttachmentByID(context.Background(), testAttachment.ID)
+	suite.ErrorIs(err, db.ErrNoEntries)
+}
+
+func (suite *PruneUnusedLocalTestSuite) TestPruneRemoteTwice() {
+	totalPruned, err := suite.manager.PruneUnusedLocalAttachments(context.Background())
+	suite.NoError(err)
+	suite.Equal(1, totalPruned)
+
+	// final prune should prune nothing, since the first prune already happened
+	totalPrunedAgain, err := suite.manager.PruneUnusedLocalAttachments(context.Background())
+	suite.NoError(err)
+	suite.Equal(0, totalPrunedAgain)
+}
+
+func (suite *PruneUnusedLocalTestSuite) TestPruneOneNonExistent() {
+	ctx := context.Background()
+	testAttachment := suite.testAttachments["local_account_1_unattached_1"]
+
+	// Delete this attachment cached on disk
+	media, err := suite.db.GetAttachmentByID(ctx, testAttachment.ID)
+	suite.NoError(err)
+	suite.True(media.Cached)
+	err = suite.storage.Delete(media.File.Path)
+	suite.NoError(err)
+
+	// Now attempt to prune for item with db entry no file
+	totalPruned, err := suite.manager.PruneUnusedLocalAttachments(ctx)
+	suite.NoError(err)
+	suite.Equal(1, totalPruned)
+}
+
+func TestPruneUnusedLocalTestSuite(t *testing.T) {
+	suite.Run(t, &PruneUnusedLocalTestSuite{})
+}

internal/media/util.go

@@ -21,6 +21,7 @@ package media
 import (
 	"errors"
 	"fmt"
+	"time"
 
 	"github.com/h2non/filetype"
 	"github.com/sirupsen/logrus"
@@ -128,3 +129,19 @@ func (l *logrusWrapper) Info(msg string, keysAndValues ...interface{}) {
 func (l *logrusWrapper) Error(err error, msg string, keysAndValues ...interface{}) {
 	logrus.Error("media manager cron logger: ", err, msg, keysAndValues)
 }
+
+func parseOlderThan(olderThanDays int) (time.Time, error) {
+	// convert days into a duration string
+	olderThanHoursString := fmt.Sprintf("%dh", olderThanDays*24)
+
+	// parse the duration string into a duration
+	olderThanHours, err := time.ParseDuration(olderThanHoursString)
+	if err != nil {
+		return time.Time{}, err
+	}
+
+	// 'subtract' that from the time now to give our threshold
+	olderThan := time.Now().Add(-olderThanHours)
+
+	return olderThan, nil
+}

internal/processing/admin/mediaprune.go

@@ -41,6 +41,15 @@ func (p *processor) MediaPrune(ctx context.Context, mediaRemoteCacheDays int) gt
 		}
 	}()
 
+	go func() {
+		pruned, err := p.mediaManager.PruneUnusedLocalAttachments(ctx)
+		if err != nil {
+			logrus.Errorf("MediaPrune: error pruning unused local cache: %s", err)
+		} else {
+			logrus.Infof("MediaPrune: pruned %d unused local cache entries", pruned)
+		}
+	}()
+
 	go func() {
 		pruned, err := p.mediaManager.PruneAllMeta(ctx)
 		if err != nil {