raft.go

/*
 * Copyright (C) 2017 Dgraph Labs, Inc. and Contributors
 *
 * This program is free software: you can redistribute it and/or modify
 * it under the terms of the GNU Affero General Public License as published by
 * the Free Software Foundation, either version 3 of the License, or
 * (at your option) any later version.
 *
 * This program is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 * GNU Affero General Public License for more details.
 *
 * You should have received a copy of the GNU Affero General Public License
 * along with this program.  If not, see <http://www.gnu.org/licenses/>.
 */

package zero

import (
	"encoding/binary"
	"errors"
	"math/rand"
	"sync"
	"time"

	"google.golang.org/grpc"

	"github.com/coreos/etcd/raft"
	"github.com/coreos/etcd/raft/raftpb"
	"github.com/dgraph-io/badger/y"
	"github.com/dgraph-io/dgraph/conn"
	"github.com/dgraph-io/dgraph/protos/intern"
	"github.com/dgraph-io/dgraph/raftwal"
	"github.com/dgraph-io/dgraph/x"
	"golang.org/x/net/context"
	"golang.org/x/net/trace"
)

type proposalCtx struct {
	ch  chan error
	ctx context.Context
}

type proposals struct {
	sync.RWMutex
	ids map[uint32]*proposalCtx
}

func (p *proposals) Store(pid uint32, pctx *proposalCtx) bool {
	if pid == 0 {
		return false
	}
	p.Lock()
	defer p.Unlock()
	if p.ids == nil {
		p.ids = make(map[uint32]*proposalCtx)
	}
	if _, has := p.ids[pid]; has {
		return false
	}
	p.ids[pid] = pctx
	return true
}

func (p *proposals) Done(pid uint32, err error) {
	p.Lock()
	defer p.Unlock()
	pd, has := p.ids[pid]
	if !has {
		return
	}
	delete(p.ids, pid)
	pd.ch <- err
}

type node struct {
	*conn.Node
	server      *Server
	ctx         context.Context
	props       proposals
	reads       map[uint64]chan uint64
	subscribers map[uint32]chan struct{}
	stop        chan struct{} // to send stop signal to Run
}

func (n *node) setRead(ch chan uint64) uint64 {
	n.Lock()
	defer n.Unlock()
	if n.reads == nil {
		n.reads = make(map[uint64]chan uint64)
	}
	for {
		ri := uint64(rand.Int63())
		if _, has := n.reads[ri]; has {
			continue
		}
		n.reads[ri] = ch
		return ri
	}
}

func (n *node) sendReadIndex(ri, id uint64) {
	n.Lock()
	ch, has := n.reads[ri]
	delete(n.reads, ri)
	n.Unlock()
	if has {
		ch <- id
	}
}

var errReadIndex = x.Errorf("cannot get linerized read (time expired or no configured leader)")

func (n *node) WaitLinearizableRead(ctx context.Context) error {
	// Read Request can get rejected then we would wait idefinitely on the channel
	// so have a timeout of 1 second.
	ctx, cancel := context.WithTimeout(ctx, time.Second)
	defer cancel()
	ch := make(chan uint64, 1)
	ri := n.setRead(ch)
	var b [8]byte
	binary.BigEndian.PutUint64(b[:], ri)
	if err := n.Raft().ReadIndex(ctx, b[:]); err != nil {
		return err
	}
	select {
	case index := <-ch:
		if index == raft.None {
			return errReadIndex
		}
		if err := n.Applied.WaitForMark(ctx, index); err != nil {
			return err
		}
		return nil
	case <-ctx.Done():
		return ctx.Err()
	}
}

func (n *node) RegisterForUpdates(ch chan struct{}) uint32 {
	n.Lock()
	defer n.Unlock()
	if n.subscribers == nil {
		n.subscribers = make(map[uint32]chan struct{})
	}
	for {
		id := rand.Uint32()
		if _, has := n.subscribers[id]; has {
			continue
		}
		n.subscribers[id] = ch
		return id
	}
}

func (n *node) Deregister(id uint32) {
	n.Lock()
	defer n.Unlock()
	delete(n.subscribers, id)
}

func (n *node) triggerUpdates() {
	n.Lock()
	defer n.Unlock()
	for _, ch := range n.subscribers {
		select {
		case ch <- struct{}{}:
		// We can ignore it and don't send a notification, because they are going to
		// read a state version after now since ch is already full.
		default:
		}
	}
}

func (n *node) AmLeader() bool {
	if n.Raft() == nil {
		return false
	}
	r := n.Raft()
	return r.Status().Lead == r.Status().ID
}

func (n *node) proposeAndWait(ctx context.Context, proposal *intern.ZeroProposal) error {
	if n.Raft() == nil {
		return x.Errorf("Raft isn't initialized yet.")
	}

	if ctx.Err() != nil {
		return ctx.Err()
	}

	che := make(chan error, 1)
	pctx := &proposalCtx{
		ch:  che,
		ctx: ctx,
	}
	for {
		id := rand.Uint32() + 1
		if n.props.Store(id, pctx) {
			proposal.Id = id
			break
		}
	}
	data, err := proposal.Marshal()
	if err != nil {
		return err
	}

	// Propose the change.
	if err := n.Raft().Propose(ctx, data); err != nil {
		return x.Wrapf(err, "While proposing")
	}

	// Wait for proposal to be applied or timeout.
	select {
	case err := <-che:
		return err
	case <-ctx.Done():
		return ctx.Err()
	}
}

var (
	errInvalidProposal     = errors.New("Invalid group proposal")
	errTabletAlreadyServed = errors.New("Tablet is already being served")
)

func newGroup() *intern.Group {
	return &intern.Group{
		Members: make(map[uint64]*intern.Member),
		Tablets: make(map[string]*intern.Tablet),
	}
}

func (n *node) applyProposal(e raftpb.Entry) (uint32, error) {
	var p intern.ZeroProposal
	// Raft commits empty entry on becoming a leader.
	if len(e.Data) == 0 {
		return p.Id, nil
	}
	if err := p.Unmarshal(e.Data); err != nil {
		return p.Id, err
	}
	if p.Id == 0 {
		return 0, errInvalidProposal
	}

	n.server.Lock()
	defer n.server.Unlock()

	state := n.server.state
	state.Counter = e.Index
	if p.MaxRaftId > 0 {
		if p.MaxRaftId <= state.MaxRaftId {
			return p.Id, errInvalidProposal
		}
		state.MaxRaftId = p.MaxRaftId
	}
	if p.Member != nil {
		m := n.server.member(p.Member.Addr)
		// Ensures that different nodes don't have same address.
		if m != nil && (m.Id != p.Member.Id || m.GroupId != p.Member.GroupId) {
			return p.Id, errInvalidAddress
		}
		if p.Member.GroupId == 0 {
			state.Zeros[p.Member.Id] = p.Member
			if p.Member.Leader {
				// Unset leader flag for other nodes, there can be only one
				// leader at a time.
				for _, m := range state.Zeros {
					if m.Id != p.Member.Id {
						m.Leader = false
					}
				}
			}
			return p.Id, nil
		}
		group := state.Groups[p.Member.GroupId]
		if group == nil {
			group = newGroup()
			state.Groups[p.Member.GroupId] = group
		}
		m, has := group.Members[p.Member.Id]
		if p.Member.AmDead {
			if has {
				delete(group.Members, p.Member.Id)
				state.Removed = append(state.Removed, m)
				conn.Get().Remove(m.Addr)
			}
			// else already removed.
			return p.Id, nil
		}
		if !has && len(group.Members) >= n.server.NumReplicas {
			// We shouldn't allow more members than the number of replicas.
			return p.Id, errInvalidProposal
		}

		// Create a connection to this server.
		go conn.Get().Connect(p.Member.Addr)

		group.Members[p.Member.Id] = p.Member
		if p.Member.Leader {
			// Unset leader flag for other nodes, there can be only one
			// leader at a time.
			for _, m := range group.Members {
				if m.Id != p.Member.Id {
					m.Leader = false
				}
			}
		}
		// On replay of logs on restart we need to set nextGroup.
		if n.server.nextGroup <= p.Member.GroupId {
			n.server.nextGroup = p.Member.GroupId + 1
		}
	}
	if p.Tablet != nil {
		if p.Tablet.GroupId == 0 {
			return p.Id, errInvalidProposal
		}
		group := state.Groups[p.Tablet.GroupId]
		if p.Tablet.Remove {
			if group != nil {
				delete(group.Tablets, p.Tablet.Predicate)
			}
			return p.Id, nil
		}
		if group == nil {
			group = newGroup()
			state.Groups[p.Tablet.GroupId] = group
		}

		// There's a edge case that we're handling.
		// Two servers ask to serve the same tablet, then we need to ensure that
		// only the first one succeeds.
		if tablet := n.server.servingTablet(p.Tablet.Predicate); tablet != nil {
			if p.Tablet.Force {
				group := state.Groups[tablet.GroupId]
				delete(group.Tablets, p.Tablet.Predicate)
			} else {
				if tablet.GroupId != p.Tablet.GroupId {
					return p.Id, errTabletAlreadyServed
				}
				// This update can come from tablet size.
				p.Tablet.ReadOnly = tablet.ReadOnly
			}
		}
		group.Tablets[p.Tablet.Predicate] = p.Tablet
	}

	if p.MaxLeaseId > state.MaxLeaseId {
		state.MaxLeaseId = p.MaxLeaseId
	} else if p.MaxTxnTs > state.MaxTxnTs {
		state.MaxTxnTs = p.MaxTxnTs
	} else if p.MaxLeaseId != 0 || p.MaxTxnTs != 0 {
		x.Printf("Could not apply lease, ignoring: proposedLease=%v maxLeaseId=%d maxTxnTs=%d",
			p, state.MaxLeaseId, state.MaxTxnTs)
	}
	if p.Txn != nil {
		n.server.orc.updateCommitStatus(e.Index, p.Txn)
	}

	return p.Id, nil
}

func (n *node) applyConfChange(e raftpb.Entry) {
	var cc raftpb.ConfChange
	cc.Unmarshal(e.Data)

	if cc.Type == raftpb.ConfChangeRemoveNode {
		n.DeletePeer(cc.NodeID)
		n.server.removeZero(cc.NodeID)
	} else if len(cc.Context) > 0 {
		var rc intern.RaftContext
		x.Check(rc.Unmarshal(cc.Context))
		n.Connect(rc.Id, rc.Addr)

		m := &intern.Member{Id: rc.Id, Addr: rc.Addr, GroupId: 0}
		n.server.storeZero(m)
	}

	cs := n.Raft().ApplyConfChange(cc)
	n.SetConfState(cs)
	n.DoneConfChange(cc.ID, nil)
	n.triggerLeaderChange()
}

func (n *node) triggerLeaderChange() {
	n.server.triggerLeaderChange()
	m := &intern.Member{Id: n.Id, Addr: n.RaftContext.Addr, Leader: n.AmLeader()}
	go n.proposeAndWait(context.Background(), &intern.ZeroProposal{Member: m})
}

func (n *node) initAndStartNode(wal *raftwal.Wal) error {
	idx, restart, err := n.InitFromWal(wal)
	n.Applied.SetDoneUntil(idx)
	x.Check(err)

	if restart {
		x.Println("Restarting node for dgraphzero")
		sp, err := n.Store.Snapshot()
		x.Checkf(err, "Unable to get existing snapshot")
		if !raft.IsEmptySnap(sp) {
			var state intern.MembershipState
			x.Check(state.Unmarshal(sp.Data))
			n.server.SetMembershipState(&state)
			for _, id := range sp.Metadata.ConfState.Nodes {
				n.Connect(id, state.Zeros[id].Addr)
			}
		}

		n.SetRaft(raft.RestartNode(n.Cfg))

	} else if len(opts.peer) > 0 {
		p := conn.Get().Connect(opts.peer)
		if p == nil {
			return errInvalidAddress
		}

		gconn := p.Get()
		c := intern.NewRaftClient(gconn)
		err = errJoinCluster
		delay := 50 * time.Millisecond
		for i := 0; i < 8 && err != nil; i++ {
			time.Sleep(delay)
			ctx, cancel := context.WithTimeout(n.ctx, time.Second)
			defer cancel()
			// JoinCluster can block idefinitely, raft ignores conf change proposal
			// if it has pending configuration.
			_, err = c.JoinCluster(ctx, n.RaftContext)
			if err == nil {
				break
			}
			if grpc.ErrorDesc(err) == conn.ErrDuplicateRaftId.Error() {
				x.Fatalf("Error while joining cluster %v", err)
			}
			x.Printf("Error while joining cluster %v\n", err)
			delay *= 2
		}
		if err != nil {
			x.Fatalf("Max retries exceeded while trying to join cluster: %v\n", err)
		}
		n.SetRaft(raft.StartNode(n.Cfg, nil))

	} else {
		data, err := n.RaftContext.Marshal()
		x.Check(err)
		peers := []raft.Peer{{ID: n.Id, Context: data}}
		n.SetRaft(raft.StartNode(n.Cfg, peers))
	}

	go n.Run()
	go n.BatchAndSendMessages()
	return err
}

func (n *node) trySnapshot() {
	existing, err := n.Store.Snapshot()
	x.Checkf(err, "Unable to get existing snapshot")
	si := existing.Metadata.Index
	idx := n.server.SyncedUntil()
	if idx <= si+1000 {
		return
	}

	data, err := n.server.MarshalMembershipState()
	x.Check(err)

	if tr, ok := trace.FromContext(n.ctx); ok {
		tr.LazyPrintf("Taking snapshot of state at watermark: %d\n", idx)
	}
	s, err := n.Store.CreateSnapshot(idx, n.ConfState(), data)
	x.Checkf(err, "While creating snapshot")
	x.Checkf(n.Store.Compact(idx), "While compacting snapshot")
	x.Check(n.Wal.StoreSnapshot(0, s))
}

func (n *node) Run() {
	var leader bool
	ticker := time.NewTicker(20 * time.Millisecond)
	defer ticker.Stop()
	rcBytes, err := n.RaftContext.Marshal()
	x.Check(err)

	closer := y.NewCloser(1)
	// We only stop runReadIndexLoop after the for loop below has finished interacting with it.
	// That way we know sending to readStateCh will not deadlock.
	defer closer.SignalAndWait()

	loop := 0
	for {
		select {
		case <-n.stop:
			n.Raft().Stop()
			return
		case <-ticker.C:
			n.Raft().Tick()

		case rd := <-n.Raft().Ready():
			for _, rs := range rd.ReadStates {
				ri := binary.BigEndian.Uint64(rs.RequestCtx)
				n.sendReadIndex(ri, rs.Index)
			}
			// First store the entries, then the hardstate and snapshot.
			x.Check(n.Wal.Store(0, rd.HardState, rd.Entries))
			x.Check(n.Wal.StoreSnapshot(0, rd.Snapshot))

			// Now store them in the in-memory store.
			n.SaveToStorage(rd.Snapshot, rd.HardState, rd.Entries)

			if !raft.IsEmptySnap(rd.Snapshot) {
				var state intern.MembershipState
				x.Check(state.Unmarshal(rd.Snapshot.Data))
				n.server.SetMembershipState(&state)
			}

			for _, entry := range rd.CommittedEntries {
				n.Applied.Begin(entry.Index)
				if entry.Type == raftpb.EntryConfChange {
					n.applyConfChange(entry)

				} else if entry.Type == raftpb.EntryNormal {
					pid, err := n.applyProposal(entry)
					if err != nil {
						x.Printf("While applying proposal: %v\n", err)
					}
					n.props.Done(pid, err)

				} else {
					x.Printf("Unhandled entry: %+v\n", entry)
				}
				n.Applied.Done(entry.Index)
			}
			loop++

			// TODO: Should we move this to the top?
			if rd.SoftState != nil {
				if rd.RaftState == raft.StateLeader && !leader {
					n.server.updateLeases()
					leader = true
				}
				// Oracle stream would close the stream once it steps down as leader
				// predicate move would cancel any in progress move on stepping down.
				n.triggerLeaderChange()
			}

			for _, msg := range rd.Messages {
				msg.Context = rcBytes
				n.Send(msg)
			}
			// Need to send membership state to dgraph nodes on leader change also.
			if rd.SoftState != nil || len(rd.CommittedEntries) > 0 {
				n.triggerUpdates()
			}
			if loop%500 == 0 {
				n.trySnapshot()
			}
			n.Raft().Advance()
		}
	}
}