Do not prune images being used by a container

Podman is not the only user of containers/storage, and as such we
cannot rely on our database as the sole source of truth when
pruning images. If images do not show as in use from Podman's
perspective, but subsequently fail to remove because they are
being used by a container, they're probably being used by Buildah
or another c/storage client.

Since the images in question are in use, we shouldn't error on
failure to prune them - we weren't supposed to prune them in the
first place.

Fixes: #3983

Signed-off-by: Matthew Heon <matthew.heon@pm.me>

1 files changed, 6 insertions, 0 deletions

diff --git a/libpod/image/prune.go b/libpod/image/prune.go
index 6ef5d321f..006cbdf22 100644
--- a/libpod/image/prune.go
+++ b/libpod/image/prune.go
@@ -4,7 +4,9 @@ import (
 	"context"
 
 	"github.com/containers/libpod/libpod/events"
+	"github.com/containers/storage"
 	"github.com/pkg/errors"
+	"github.com/sirupsen/logrus"
 )
 
 // GetPruneImages returns a slice of images that have no names/unused
@@ -44,6 +46,10 @@ func (ir *Runtime) PruneImages(ctx context.Context, all bool) ([]string, error)
 	}
 	for _, p := range pruneImages {
 		if err := p.Remove(ctx, true); err != nil {
+			if errors.Cause(err) == storage.ErrImageUsedByContainer {
+				logrus.Warnf("Failed to prune image %s as it is in use: %v", p.ID(), err)
+				continue
+			}
 			return nil, errors.Wrap(err, "failed to prune image")
 		}
 		defer p.newImageEvent(events.Prune)