etcdserver: raise alarm on cluster corruption

Fixes etcd-io#7125

etcdserver/apply.go

@@ -577,9 +577,11 @@ func (a *applierV3backend) Alarm(ar *pb.AlarmRequest) (*pb.AlarmResponse, error)
 			break
 		}
 
+		plog.Warningf("alarm %v raised by peer %s", m.Alarm, types.ID(m.MemberID))
 		switch m.Alarm {
+		case pb.AlarmType_CORRUPT:
+			a.s.applyV3 = newApplierV3Corrupt(a)
 		case pb.AlarmType_NOSPACE:
-			plog.Warningf("alarm raised %+v", m)
 			a.s.applyV3 = newApplierV3Capped(a)
 		default:
 			plog.Errorf("unimplemented alarm activation (%+v)", m)
@@ -596,7 +598,8 @@ func (a *applierV3backend) Alarm(ar *pb.AlarmRequest) (*pb.AlarmResponse, error)
 		}
 
 		switch m.Alarm {
-		case pb.AlarmType_NOSPACE:
+		case pb.AlarmType_NOSPACE, pb.AlarmType_CORRUPT:
+			// TODO: check kv hash before deactivating CORRUPT?
 			plog.Infof("alarm disarmed %+v", ar)
 			a.s.applyV3 = a.s.newApplierV3()
 		default:

etcdserver/config.go

@@ -66,6 +66,8 @@ type ServerConfig struct {
 	ClientCertAuthEnabled bool
 
 	AuthToken string
+
+	CorruptCheckTime time.Duration
 }
 
 // VerifyBootstrap sanity-checks the initial config for bootstrap case

etcdserver/corrupt.go

@@ -0,0 +1,182 @@
+// Copyright 2017 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 etcdserver
+
+import (
+	"time"
+
+	"github.com/coreos/etcd/clientv3"
+	pb "github.com/coreos/etcd/etcdserver/etcdserverpb"
+	"github.com/coreos/etcd/mvcc"
+	"github.com/coreos/etcd/pkg/types"
+
+	"golang.org/x/net/context"
+)
+
+func (s *EtcdServer) monitorKVHash() {
+	t := s.Cfg.CorruptCheckTime
+	if t == 0 {
+		return
+	}
+	plog.Infof("enabled corruption checking with %s interval", t)
+	for {
+		select {
+		case <-s.stopping:
+			return
+		case <-time.After(t):
+		}
+		if !s.isLeader() {
+			continue
+		}
+		if err := s.checkHashKV(); err != nil {
+			plog.Debugf("check hash kv failed %v", err)
+		}
+	}
+}
+
+func (s *EtcdServer) checkHashKV() error {
+	h, rev, crev, err := s.kv.HashByRev(0)
+	if err != nil {
+		plog.Fatalf("failed to hash kv store (%v)", err)
+	}
+	resps := []*clientv3.HashKVResponse{}
+	for _, m := range s.cluster.Members() {
+		if m.ID == s.ID() {
+			continue
+		}
+
+		cli, cerr := clientv3.New(clientv3.Config{Endpoints: m.PeerURLs})
+		if cerr != nil {
+			continue
+		}
+
+		respsLen := len(resps)
+		for _, c := range cli.Endpoints() {
+			ctx, cancel := context.WithTimeout(context.Background(), s.Cfg.ReqTimeout())
+			resp, herr := cli.HashKV(ctx, c, rev)
+			cancel()
+			if herr == nil {
+				cerr = herr
+				resps = append(resps, resp)
+				break
+			}
+		}
+		cli.Close()
+
+		if respsLen == len(resps) {
+			plog.Warningf("failed to hash kv for peer %s (%v)", types.ID(m.ID), cerr)
+		}
+	}
+
+	ctx, cancel := context.WithTimeout(context.Background(), s.Cfg.ReqTimeout())
+	err = s.linearizableReadNotify(ctx)
+	cancel()
+	if err != nil {
+		return err
+	}
+
+	h2, rev2, crev2, err := s.kv.HashByRev(0)
+	if err != nil {
+		plog.Warningf("failed to hash kv store (%v)", err)
+		return err
+	}
+
+	alarmed := false
+	mismatch := func(id uint64) {
+		if alarmed {
+			return
+		}
+		alarmed = true
+		a := &pb.AlarmRequest{
+			MemberID: uint64(id),
+			Action:   pb.AlarmRequest_ACTIVATE,
+			Alarm:    pb.AlarmType_CORRUPT,
+		}
+		s.goAttach(func() {
+			s.raftRequest(s.ctx, pb.InternalRaftRequest{Alarm: a})
+		})
+	}
+
+	if h2 != h && rev2 == rev && crev == crev2 {
+		plog.Warningf("mismatched hashes %d and %d for revision %d", h, h2, rev)
+		mismatch(uint64(s.ID()))
+	}
+
+	for _, resp := range resps {
+		id := resp.Header.MemberId
+		if resp.Header.Revision > rev2 {
+			plog.Warningf(
+				"revision %d from member %v, expected at most %d",
+				resp.Header.Revision,
+				types.ID(id),
+				rev2)
+			mismatch(id)
+		}
+		if resp.CompactRevision > crev2 {
+			plog.Warningf(
+				"compact revision %d from member %v, expected at most %d",
+				resp.CompactRevision,
+				types.ID(id),
+				crev2,
+			)
+			mismatch(id)
+		}
+		if resp.CompactRevision == crev && resp.Hash != h {
+			plog.Warningf(
+				"hash %d at revision %d from member %v, expected hash %d",
+				resp.Hash,
+				rev,
+				types.ID(id),
+				h,
+			)
+			mismatch(id)
+		}
+	}
+	return nil
+}
+
+type applierV3Corrupt struct {
+	applierV3
+}
+
+func newApplierV3Corrupt(a applierV3) *applierV3Corrupt { return &applierV3Corrupt{a} }
+
+func (a *applierV3Corrupt) Put(txn mvcc.TxnWrite, p *pb.PutRequest) (*pb.PutResponse, error) {
+	return nil, ErrCorrupt
+}
+
+func (a *applierV3Corrupt) Range(txn mvcc.TxnRead, p *pb.RangeRequest) (*pb.RangeResponse, error) {
+	return nil, ErrCorrupt
+}
+
+func (a *applierV3Corrupt) DeleteRange(txn mvcc.TxnWrite, p *pb.DeleteRangeRequest) (*pb.DeleteRangeResponse, error) {
+	return nil, ErrCorrupt
+}
+
+func (a *applierV3Corrupt) Txn(rt *pb.TxnRequest) (*pb.TxnResponse, error) {
+	return nil, ErrCorrupt
+}
+
+func (a *applierV3Corrupt) Compaction(compaction *pb.CompactionRequest) (*pb.CompactionResponse, <-chan struct{}, error) {
+	return nil, nil, ErrCorrupt
+}
+
+func (a *applierV3Corrupt) LeaseGrant(lc *pb.LeaseGrantRequest) (*pb.LeaseGrantResponse, error) {
+	return nil, ErrCorrupt
+}
+
+func (a *applierV3Corrupt) LeaseRevoke(lc *pb.LeaseRevokeRequest) (*pb.LeaseRevokeResponse, error) {
+	return nil, ErrCorrupt
+}

etcdserver/errors.go

@@ -35,6 +35,7 @@ var (
 	ErrTooManyRequests            = errors.New("etcdserver: too many requests")
 	ErrUnhealthy                  = errors.New("etcdserver: unhealthy cluster")
 	ErrKeyNotFound                = errors.New("etcdserver: key not found")
+	ErrCorrupt                    = errors.New("etcdserver: corrupt cluster")
 )
 
 type DiscoveryError struct {

etcdserver/server.go

@@ -524,6 +524,7 @@ func (s *EtcdServer) Start() {
 	s.goAttach(func() { monitorFileDescriptor(s.stopping) })
 	s.goAttach(s.monitorVersions)
 	s.goAttach(s.linearizableReadLoop)
+	s.goAttach(s.monitorKVHash)
 }
 
 // start prepares and starts server in a new goroutine. It is no longer safe to
@@ -1628,6 +1629,9 @@ func (s *EtcdServer) restoreAlarms() error {
 	if len(as.Get(pb.AlarmType_NOSPACE)) > 0 {
 		s.applyV3 = newApplierV3Capped(s.applyV3)
 	}
+	if len(as.Get(pb.AlarmType_CORRUPT)) > 0 {
+		s.applyV3 = newApplierV3Corrupt(s.applyV3)
+	}
 	return nil
 }