fix the data inconsistency issue by adding a txPostLockHook into the backend

Previously the SetConsistentIndex() is called during the apply workflow,
but it's outside the db transaction. If a commit happens between SetConsistentIndex
and the following apply workflow, and etcd crashes for whatever reason right
after the commit, then etcd commits an incomplete transaction to db.
Eventually etcd runs into the data inconsistency issue.

In this commit, we move the SetConsistentIndex into a txPostLockHook, so
it will be executed inside the transaction lock.

server/etcdserver/server_test.go

@@ -686,9 +686,7 @@ func TestApplyConfigChangeUpdatesConsistIndex(t *testing.T) {
 
 	_, appliedi, _ := srv.apply(ents, &raftpb.ConfState{})
 	consistIndex := srv.consistIndex.ConsistentIndex()
-	if consistIndex != appliedi {
-		t.Fatalf("consistIndex = %v, want %v", consistIndex, appliedi)
-	}
+	assert.Equal(t, uint64(2), appliedi)
 
 	t.Run("verify-backend", func(t *testing.T) {
 		tx := be.BatchTx()
@@ -697,9 +695,8 @@ func TestApplyConfigChangeUpdatesConsistIndex(t *testing.T) {
 		srv.beHooks.OnPreCommitUnsafe(tx)
 		assert.Equal(t, raftpb.ConfState{Voters: []uint64{2}}, *membership.UnsafeConfStateFromBackend(lg, tx))
 	})
-	rindex, rterm := cindex.ReadConsistentIndex(be.BatchTx())
+	rindex, _ := cindex.ReadConsistentIndex(be.ReadTx())
 	assert.Equal(t, consistIndex, rindex)
-	assert.Equal(t, uint64(4), rterm)
 }
 
 func realisticRaftNode(lg *zap.Logger) *raftNode {