etcd/raft/rafttest/interaction_env_handler_add_nodes.go
Tobias Schottdorf e8090e57a2 raft/rafttest: introduce datadriven testing
It has often been tedious to test the interactions between multi-member
Raft groups, especially when many steps were required to reach a certain
scenario. Often, this boilerplate was as boring as it is hard to write
and hard to maintain, making it attractive to resort to shortcuts
whenever possible, which in turn tended to undercut how meaningful and
maintainable the tests ended up being - that is, if the tests were even
written, which sometimes they weren't.

This change introduces a datadriven framework specifically for testing
deterministically the interaction between multiple members of a raft group
with the goal of reducing the friction for writing these tests to near
zero.

In the near term, this will be used to add thorough testing for joint
consensus (which is already available today, but wildly undertested),
but just converting an existing test into this framework has shown that
the concise representation and built-in inspection of log messages
highlights unexpected behavior much more readily than the previous unit
tests did (the test in question is `snapshot_succeed_via_app_resp`; the
reader is invited to compare the old and new version of it).

The main building block is `InteractionEnv`, which holds on to the state
of the whole system and exposes various relevant methods for
manipulating it, including but not limited to adding nodes, delivering
and dropping messages, and proposing configuration changes. All of this
is extensible so that in the future I hope to use it to explore the
phenomena discussed in

https://github.com/etcd-io/etcd/issues/7625#issuecomment-488798263

which requires injecting appropriate "crash points" in the Ready
handling loop. Discussions of the "what if X happened in state Y"
can quickly be made concrete by "scripting up an interaction test".

Additionally, this framework is intentionally not kept internal to the
raft package.. Though this is in its infancy, a goal is that it should
be possible for a suite of interaction tests to allow applications to
validate that their Storage implementation behaves accordingly, simply
by running a raft-provided interaction suite against their Storage.
2019-08-12 11:13:51 +02:00

137 lines
4.0 KiB
Go

// Copyright 2019 The etcd Authors
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
package rafttest
import (
"errors"
"fmt"
"reflect"
"testing"
"github.com/cockroachdb/datadriven"
"go.etcd.io/etcd/raft"
pb "go.etcd.io/etcd/raft/raftpb"
)
func (env *InteractionEnv) handleAddNodes(t *testing.T, d datadriven.TestData) error {
n := firstAsInt(t, d)
var snap pb.Snapshot
for _, arg := range d.CmdArgs[1:] {
for i := range arg.Vals {
switch arg.Key {
case "voters":
var id uint64
arg.Scan(t, i, &id)
snap.Metadata.ConfState.Voters = append(snap.Metadata.ConfState.Voters, id)
case "learners":
var id uint64
arg.Scan(t, i, &id)
snap.Metadata.ConfState.Learners = append(snap.Metadata.ConfState.Learners, id)
case "index":
arg.Scan(t, i, &snap.Metadata.Index)
case "content":
arg.Scan(t, i, &snap.Data)
}
}
}
return env.AddNodes(n, snap)
}
type snapOverrideStorage struct {
Storage
snapshotOverride func() (pb.Snapshot, error)
}
func (s snapOverrideStorage) Snapshot() (pb.Snapshot, error) {
if s.snapshotOverride != nil {
return s.snapshotOverride()
}
return s.Storage.Snapshot()
}
var _ raft.Storage = snapOverrideStorage{}
// AddNodes adds n new nodes initializes from the given snapshot (which may be
// empty). They will be assigned consecutive IDs.
func (env *InteractionEnv) AddNodes(n int, snap pb.Snapshot) error {
bootstrap := !reflect.DeepEqual(snap, pb.Snapshot{})
for i := 0; i < n; i++ {
id := uint64(1 + len(env.Nodes))
s := snapOverrideStorage{
Storage: raft.NewMemoryStorage(),
// When you ask for a snapshot, you get the most recent snapshot.
//
// TODO(tbg): this is sort of clunky, but MemoryStorage itself will
// give you some fixed snapshot and also the snapshot changes
// whenever you compact the logs and vice versa, so it's all a bit
// awkward to use.
snapshotOverride: func() (pb.Snapshot, error) {
snaps := env.Nodes[int(id-1)].History
return snaps[len(snaps)-1], nil
},
}
if bootstrap {
// NB: we could make this work with 1, but MemoryStorage just
// doesn't play well with that and it's not a loss of generality.
if snap.Metadata.Index <= 1 {
return errors.New("index must be specified as > 1 due to bootstrap")
}
snap.Metadata.Term = 1
if err := s.ApplySnapshot(snap); err != nil {
return err
}
fi, err := s.FirstIndex()
if err != nil {
return err
}
// At the time of writing and for *MemoryStorage, applying a
// snapshot also truncates appropriately, but this would change with
// other storage engines potentially.
if exp := snap.Metadata.Index + 1; fi != exp {
return fmt.Errorf("failed to establish first index %d; got %d", exp, fi)
}
}
cfg := defaultRaftConfig(id, snap.Metadata.Index, s)
if env.Options.OnConfig != nil {
env.Options.OnConfig(cfg)
if cfg.ID != id {
// This could be supported but then we need to do more work
// translating back and forth -- not worth it.
return errors.New("OnConfig must not change the ID")
}
}
if cfg.Logger != nil {
return errors.New("OnConfig must not set Logger")
}
cfg.Logger = env.Output
rn, err := raft.NewRawNode(cfg)
if err != nil {
return err
}
node := Node{
RawNode: rn,
// TODO(tbg): allow a more general Storage, as long as it also allows
// us to apply snapshots, append entries, and update the HardState.
Storage: s,
Config: cfg,
History: []pb.Snapshot{snap},
}
env.Nodes = append(env.Nodes, node)
}
return nil
}