S3: do not set storage class for metadata when using archive storage

changelog/unreleased/issue-4583

@@ -0,0 +1,9 @@
+Bugfix: Ignoring the s3.storage-class option for metadata when archive tier is specified
+
+Restic now will save snapshot metadata to non-archive storage tier whatsoever,
+this will help avoid issues when data is being saved to archive storage class.
+It is not providing any support for cold storages in restic, 
+only saving users from making backups unusable.
+
+https://github.com/restic/restic/issues/4583
+https://github.com/restic/restic/issues/3202

internal/backend/s3/s3.go

@@ -325,12 +325,22 @@ func (be *Backend) Path() string {
 	return be.cfg.Prefix
 }
 
+// useStorageClass returns whether file should be saved in the provided Storage Class
+func (be *Backend) useStorageClass(h backend.Handle) bool {
+	var notArchiveClass bool = be.cfg.StorageClass != "GLACIER" && be.cfg.StorageClass != "DEEP_ARCHIVE"
+	isDataFile := h.Type == backend.PackFile && !h.IsMetadata
+	return isDataFile || notArchiveClass
+}
+
 // Save stores data in the backend at the handle.
 func (be *Backend) Save(ctx context.Context, h backend.Handle, rd backend.RewindReader) error {
 	objName := be.Filename(h)
 
-	opts := minio.PutObjectOptions{StorageClass: be.cfg.StorageClass}
-	opts.ContentType = "application/octet-stream"
+	opts := minio.PutObjectOptions{ContentType: "application/octet-stream"}
+
+	if be.useStorageClass(h) {
+		opts.StorageClass = be.cfg.StorageClass
+	}
 	// the only option with the high-level api is to let the library handle the checksum computation
 	opts.SendContentMd5 = true
 	// only use multipart uploads for very large files