repoclean: Log warning for missed repositories

For each repository that is not know to praefect but
that exists on one of the gitaly nodes in the cluster
the warning message will be logged with the info about
virtual storage and gitaly node as well as relative path
to the missing repository. Based on this info admin could
remove the repository if it is a garbage or propagate it
to the praefect, so it becomes available for serving.

Part of: https://gitlab.com/gitlab-org/gitaly/-/issues/3719

2 files changed, 81 insertions, 0 deletions

diff --git a/internal/praefect/repocleaner/action_log.go b/internal/praefect/repocleaner/action_log.go
new file mode 100644
index 000000000..934d69c49
--- /dev/null
+++ b/internal/praefect/repocleaner/action_log.go
@@ -0,0 +1,33 @@
+package repocleaner
+
+import (
+	"context"
+
+	"github.com/sirupsen/logrus"
+	"gitlab.com/gitlab-org/gitaly/v14/internal/praefect/datastore"
+)
+
+// LogWarnAction is an implementation of the Action interface that allows to log a warning message
+// for the repositories that are not known for the praefect.
+type LogWarnAction struct {
+	logger logrus.FieldLogger
+}
+
+// NewLogWarnAction return new instance of the LogWarnAction.
+func NewLogWarnAction(logger logrus.FieldLogger) *LogWarnAction {
+	return &LogWarnAction{
+		logger: logger.WithField("component", "repocleaner.log_warn_action"),
+	}
+}
+
+// Perform logs a warning for each repository that is not known to praefect.
+func (al LogWarnAction) Perform(_ context.Context, notExisting []datastore.RepositoryClusterPath) error {
+	for _, entry := range notExisting {
+		al.logger.WithFields(logrus.Fields{
+			"virtual_storage":       entry.VirtualStorage,
+			"storage":               entry.Storage,
+			"relative_replica_path": entry.RelativeReplicaPath,
+		}).Warn("repository is not managed by praefect")
+	}
+	return nil
+}
diff --git a/internal/praefect/repocleaner/action_log_test.go b/internal/praefect/repocleaner/action_log_test.go
new file mode 100644
index 000000000..1c7af50b2
--- /dev/null
+++ b/internal/praefect/repocleaner/action_log_test.go
@@ -0,0 +1,48 @@
+package repocleaner
+
+import (
+	"context"
+	"testing"
+
+	"github.com/sirupsen/logrus"
+	"github.com/sirupsen/logrus/hooks/test"
+	"github.com/stretchr/testify/require"
+	"gitlab.com/gitlab-org/gitaly/v14/internal/praefect/datastore"
+)
+
+func TestLogWarnAction_Perform(t *testing.T) {
+	logger, hook := test.NewNullLogger()
+	action := NewLogWarnAction(logger)
+	err := action.Perform(context.TODO(), []datastore.RepositoryClusterPath{
+		{ClusterPath: datastore.ClusterPath{VirtualStorage: "vs1", Storage: "g1"}, RelativeReplicaPath: "p/1"},
+		{ClusterPath: datastore.ClusterPath{VirtualStorage: "vs2", Storage: "g2"}, RelativeReplicaPath: "p/2"},
+	})
+	require.NoError(t, err)
+	require.Len(t, hook.AllEntries(), 2)
+
+	exp := []map[string]interface{}{{
+		"Data": logrus.Fields{
+			"component":             "repocleaner.log_warn_action",
+			"virtual_storage":       "vs1",
+			"storage":               "g1",
+			"relative_replica_path": "p/1",
+		},
+		"Message": "repository is not managed by praefect",
+	}, {
+		"Data": logrus.Fields{
+			"component":             "repocleaner.log_warn_action",
+			"virtual_storage":       "vs2",
+			"storage":               "g2",
+			"relative_replica_path": "p/2",
+		},
+		"Message": "repository is not managed by praefect",
+	}}
+
+	require.ElementsMatch(t, exp, []map[string]interface{}{{
+		"Data":    hook.AllEntries()[0].Data,
+		"Message": hook.AllEntries()[0].Message,
+	}, {
+		"Data":    hook.AllEntries()[1].Data,
+		"Message": hook.AllEntries()[1].Message,
+	}})
+}