X-Git-Url: https://git.arvados.org/arvados.git/blobdiff_plain/139200027a3192260b5ea7c2d0c93a8eb5f8eb7e..1767cc94ddd427c6610c82e1b27f6a9f6793b39a:/services/keep-balance/collection.go

diff --git a/services/keep-balance/collection.go b/services/keep-balance/collection.go
index e6a1f08cf1..f12522fb6c 100644
--- a/services/keep-balance/collection.go
+++ b/services/keep-balance/collection.go
@@ -22,21 +22,31 @@ func countCollections(c *arvados.Client, params arvados.ResourceListParams) (int
 // The progress function is called periodically with done (number of
 // times f has been called) and total (number of times f is expected
 // to be called).
-func EachCollection(c *arvados.Client, f func(arvados.Collection) error, progress func(done, total int)) error {
+//
+// If pageSize > 0 it is used as the maximum page size in each API
+// call; otherwise the maximum allowed page size is requested.
+func EachCollection(c *arvados.Client, pageSize int, f func(arvados.Collection) error, progress func(done, total int)) error {
 	if progress == nil {
 		progress = func(_, _ int) {}
 	}
 
-	expectCount, err := countCollections(c, arvados.ResourceListParams{})
+	expectCount, err := countCollections(c, arvados.ResourceListParams{
+		IncludeTrash: true,
+	})
 	if err != nil {
 		return err
 	}
 
-	limit := 1000
+	limit := pageSize
+	if limit <= 0 {
+		// Use the maximum page size the server allows
+		limit = 1<<31 - 1
+	}
 	params := arvados.ResourceListParams{
-		Limit:  &limit,
-		Order:  "modified_at, uuid",
-		Select: []string{"uuid", "manifest_text", "modified_at", "portable_data_hash", "replication_desired"},
+		Limit:        &limit,
+		Order:        "modified_at, uuid",
+		Select:       []string{"uuid", "manifest_text", "modified_at", "portable_data_hash", "replication_desired"},
+		IncludeTrash: true,
 	}
 	var last arvados.Collection
 	var filterTime time.Time
@@ -82,10 +92,13 @@ func EachCollection(c *arvados.Client, f func(arvados.Collection) error, progres
 	}
 	progress(callCount, expectCount)
 
-	if checkCount, err := countCollections(c, arvados.ResourceListParams{Filters: []arvados.Filter{{
-		Attr:     "modified_at",
-		Operator: "<=",
-		Operand:  filterTime}}}); err != nil {
+	if checkCount, err := countCollections(c, arvados.ResourceListParams{
+		Filters: []arvados.Filter{{
+			Attr:     "modified_at",
+			Operator: "<=",
+			Operand:  filterTime}},
+		IncludeTrash: true,
+	}); err != nil {
 		return err
 	} else if callCount < checkCount {
 		return fmt.Errorf("Retrieved %d collections with modtime <= T=%q, but server now reports there are %d collections with modtime <= T", callCount, filterTime, checkCount)