5985667653

Committed 26 Aug 2023 03:46PM UTC coverage: 66.655% (-0.5%) from 67.169%

Build Type

push

Committed by

Commit Message

chore(deps): bump github.com/apache/thrift from 0.12.0 to 0.13.0 (#8982)

* Bumps github.com/apache/thrift from 0.12.0 to 0.13.0

Signed-off-by: dependabot[bot] <support@github.com>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>

Run Details

58142 of 87228 relevant lines covered (66.66%)

2209533.57 hits per line

Source File
Press 'n' to go to next uncovered line, 'b' for previous

75.79

/dgraph/cmd/zero/tablet.go

/*
 * Copyright 2017-2023 Dgraph Labs, Inc. and Contributors
 *
 * 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 zero

import (
        "context"
        "fmt"
        "sort"
        "time"

        humanize "github.com/dustin/go-humanize"
        "github.com/golang/glog"
        "github.com/pkg/errors"
        otrace "go.opencensus.io/trace"

        "github.com/dgraph-io/dgraph/protos/pb"
        "github.com/dgraph-io/dgraph/x"
)

const (
        predicateMoveTimeout = 120 * time.Minute
)

/*
Steps to move predicate p from g1 to g2.
Design change:
• If you’re not the leader, don’t talk to zero.
• Let the leader send you updates via proposals.

Move:
• Dgraph zero would decide that G1 should not serve P, G2 should serve it.
• Zero would propose that G1 is read-only for predicate P. This would propagate to the cluster.

• Zero would tell G1 to move P to G2 (Endpoint: Zero → G1)

This would trigger G1 to get latest state. Wait for it.
• G1 would propose this state to it’s followers.
• G1 after proposing would do a call to G2, and start streaming.
• Before G2 starts accepting, it should delete any current keys for P.
• It should tell Zero whether it succeeded or failed. (Endpoint: G1 → Zero)

• Zero would then propose that G2 is serving P (or G1 is, if fail above) P would RW.
• G1 gets this, G2 gets this.
• Both propagate this to their followers.

*/

// TODO: Have a event log for everything.
func (s *Server) rebalanceTablets() {
        ticker := time.NewTicker(opts.rebalanceInterval)
        for range ticker.C {
                predicate, srcGroup, dstGroup := s.chooseTablet()
                if len(predicate) == 0 {
                        continue
                }
                if err := s.movePredicate(predicate, srcGroup, dstGroup); err != nil {
                        glog.Errorln(err)
                }
        }
}

// MoveTablet can be used to move a tablet to a specific group.
// It takes in tablet and destination group as argument.
// It returns a *pb.Status to be used by the `/moveTablet` HTTP handler in Zero.
func (s *Server) MoveTablet(ctx context.Context, req *pb.MoveTabletRequest) (*pb.Status, error) {
        if !s.Node.AmLeader() {
                return &pb.Status{Code: 1, Msg: x.Error}, errNotLeader
        }

        knownGroups := s.KnownGroups()
        var isKnown bool
        for _, grp := range knownGroups {
                if grp == req.DstGroup {
                        isKnown = true
                        break
                }
        }
        if !isKnown {
                return &pb.Status{Code: 1, Msg: x.ErrorInvalidRequest},
                        fmt.Errorf("group: [%d] is not a known group", req.DstGroup)
        }

        tablet := x.NamespaceAttr(req.Namespace, req.Tablet)
        tab := s.ServingTablet(tablet)
        if tab == nil {
                return &pb.Status{Code: 1, Msg: x.ErrorInvalidRequest},
                        fmt.Errorf("namespace: %d. No tablet found for: %s", req.Namespace, req.Tablet)
        }

        srcGroup := tab.GroupId
        if srcGroup == req.DstGroup {
                return &pb.Status{Code: 1, Msg: x.ErrorInvalidRequest},
                        fmt.Errorf("namespace: %d. Tablet: [%s] is already being served by group: [%d]",
                                req.Namespace, req.Tablet, srcGroup)
        }

        if err := s.movePredicate(tablet, srcGroup, req.DstGroup); err != nil {
                glog.Errorf("namespace: %d. While moving predicate %s from %d -> %d. Error: %v",
                        req.Namespace, req.Tablet, srcGroup, req.DstGroup, err)
                return &pb.Status{Code: 1, Msg: x.Error}, err
        }

        return &pb.Status{Code: 0, Msg: fmt.Sprintf("namespace: %d. "+
                "Predicate: [%s] moved from group [%d] to [%d]", req.Namespace, req.Tablet, srcGroup,
                req.DstGroup)}, nil
}

// movePredicate is the main entry point for move predicate logic. This Zero must remain the leader
// for the entire duration of predicate move. If this Zero stops being the leader, the final
// proposal of reassigning the tablet to the destination would fail automatically.
func (s *Server) movePredicate(predicate string, srcGroup, dstGroup uint32) error {
        s.moveOngoing <- struct{}{}
        defer func() {
                <-s.moveOngoing
        }()

        ctx, cancel := context.WithTimeout(context.Background(), predicateMoveTimeout)
        defer cancel()

        ctx, span := otrace.StartSpan(ctx, "Zero.MovePredicate")
        defer span.End()

        // Ensure that reserved predicates cannot be moved.
        if x.IsReservedPredicate(predicate) {
                return errors.Errorf("Unable to move reserved predicate %s", predicate)
        }

        // Ensure that I'm connected to the rest of the Zero group, and am the leader.
        if _, err := s.latestMembershipState(ctx); err != nil {
                return errors.Wrapf(err, "unable to reach quorum")
        }
        if !s.Node.AmLeader() {
                return errors.Errorf("I am not the Zero leader")
        }
        tab := s.ServingTablet(predicate)
        if tab == nil {
                return errors.Errorf("Tablet to be moved: [%v] is not being served", predicate)
        }
        msg := fmt.Sprintf("Going to move predicate: [%v], size: [ondisk: %v, uncompressed: %v]"+
                " from group %d to %d\n", predicate, humanize.IBytes(uint64(tab.OnDiskBytes)),
                humanize.IBytes(uint64(tab.UncompressedBytes)), srcGroup, dstGroup)
        glog.Info(msg)
        span.Annotate([]otrace.Attribute{otrace.StringAttribute("tablet", predicate)}, msg)

        // Block all commits on this predicate. Keep them blocked until we return from this function.
        unblock := s.blockTablet(predicate)
        defer unblock()

        // Get a new timestamp, beyond which we are sure that no new txns would be committed for this
        // predicate. Source Alpha leader must reach this timestamp before streaming the data.
        ids, err := s.Timestamps(ctx, &pb.Num{Val: 1})
        if err != nil || ids.StartId == 0 {
                return errors.Wrapf(err, "while leasing txn timestamp. Id: %+v", ids)
        }

        // Get connection to leader of source group.
        pl := s.Leader(srcGroup)
        if pl == nil {
                return errors.Errorf("No healthy connection found to leader of group %d", srcGroup)
        }
        wc := pb.NewWorkerClient(pl.Get())
        in := &pb.MovePredicatePayload{
                Predicate: predicate,
                SourceGid: srcGroup,
                DestGid:   dstGroup,
                TxnTs:     ids.StartId,
        }
        span.Annotatef(nil, "Starting move: %+v", in)
        glog.Infof("Starting move: %+v", in)
        if _, err := wc.MovePredicate(ctx, in); err != nil {
                return errors.Wrapf(err, "while calling MovePredicate")
        }

        p := &pb.ZeroProposal{}
        p.Tablet = &pb.Tablet{
                GroupId:           dstGroup,
                Predicate:         predicate,
                OnDiskBytes:       tab.OnDiskBytes,
                UncompressedBytes: tab.UncompressedBytes,
                Force:             true,
                MoveTs:            in.TxnTs,
        }
        msg = fmt.Sprintf("Move at Alpha done. Now proposing: %+v", p)
        span.Annotate(nil, msg)
        glog.Info(msg)
        if err := s.Node.proposeAndWait(ctx, p); err != nil {
                return errors.Wrapf(err, "while proposing tablet reassignment. Proposal: %+v", p)
        }
        msg = fmt.Sprintf("Predicate move done for: [%v] from group %d to %d\n",
                predicate, srcGroup, dstGroup)
        glog.Info(msg)
        span.Annotate(nil, msg)

        // Now that the move has happened, we can delete the predicate from the source group. But before
        // doing that, we should ensure the source group understands that the predicate is now being
        // served by the destination group. For that, we pass in the expected checksum for the source
        // group. Only once the source group membership checksum matches, would the source group delete
        // the predicate. This ensures that it does not service any transaction after deletion of data.
        checksums := s.groupChecksums()
        in.ExpectedChecksum = checksums[in.SourceGid]
        in.DestGid = 0 // Indicates deletion of predicate in the source group.
        if _, err := wc.MovePredicate(ctx, in); err != nil {
                msg = fmt.Sprintf("While deleting predicate [%v] in group %d. Error: %v",
                        in.Predicate, in.SourceGid, err)
                span.Annotate(nil, msg)
                glog.Warningf(msg)
        } else {
                msg = fmt.Sprintf("Deleted predicate %v in group %d", in.Predicate, in.SourceGid)
                span.Annotate(nil, msg)
                glog.V(1).Infof(msg)
        }
        return nil
}

func (s *Server) chooseTablet() (predicate string, srcGroup uint32, dstGroup uint32) {
        s.RLock()
        defer s.RUnlock()
        if s.state == nil {
                return
        }
        numGroups := len(s.state.Groups)
        if !s.Node.AmLeader() || numGroups <= 1 {
                return
        }

        // Sort all groups by their sizes.
        type kv struct {
                gid  uint32
                size int64 // in bytes
        }
        var groups []kv
        for k, v := range s.state.Groups {
                space := int64(0)
                for _, tab := range v.Tablets {
                        space += tab.OnDiskBytes
                }
                groups = append(groups, kv{k, space})
        }
        sort.Slice(groups, func(i, j int) bool {
                return groups[i].size < groups[j].size
        })

        glog.Infof("\n\nGroups sorted by size: %+v\n\n", groups)
        for lastGroup := numGroups - 1; lastGroup > 0; lastGroup-- {
                srcGroup = groups[lastGroup].gid
                dstGroup = groups[0].gid
                sizeDiff := groups[lastGroup].size - groups[0].size
                glog.Infof("size_diff %v\n", sizeDiff)
                // Don't move a node unless you receive atleast one update regarding tablet size.
                // Tablet size would have come up with leader update.
                if !s.hasLeader(dstGroup) {
                        return
                }
                // We move the predicate only if the difference between size of both machines is
                // atleast 10% of dst group.
                if float64(sizeDiff) < 0.1*float64(groups[0].size) {
                        continue
                }

                // Try to find a predicate which we can move.
                size := int64(0)
                group := s.state.Groups[srcGroup]
                for _, tab := range group.Tablets {
                        // Reserved predicates should always be in group 1 so do not re-balance them.
                        if x.IsReservedPredicate(tab.Predicate) {
                                continue
                        }

                        // Finds a tablet as big a possible such that on moving it dstGroup's size is
                        // less than or equal to srcGroup.
                        if tab.OnDiskBytes <= sizeDiff/2 && tab.OnDiskBytes > size {
                                predicate = tab.Predicate
                                size = tab.OnDiskBytes
                        }
                }
                if len(predicate) > 0 {
                        return
                }
        }
        return
}

1	/*
2	* Copyright 2017-2023 Dgraph Labs, Inc. and Contributors
3	*
4	* Licensed under the Apache License, Version 2.0 (the "License");
5	* you may not use this file except in compliance with the License.
6	* You may obtain a copy of the License at
7	*
8	* http://www.apache.org/licenses/LICENSE-2.0
9	*
10	* Unless required by applicable law or agreed to in writing, software
11	* distributed under the License is distributed on an "AS IS" BASIS,
12	* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
13	* See the License for the specific language governing permissions and
14	* limitations under the License.
15	*/
16
17	package zero
18
19	import (
20	"context"
21	"fmt"
22	"sort"
23	"time"
24
25	humanize "github.com/dustin/go-humanize"
26	"github.com/golang/glog"
27	"github.com/pkg/errors"
28	otrace "go.opencensus.io/trace"
29
30	"github.com/dgraph-io/dgraph/protos/pb"
31	"github.com/dgraph-io/dgraph/x"
32	)
33
34	const (
35	predicateMoveTimeout = 120 * time.Minute
36	)
37
38	/*
39	Steps to move predicate p from g1 to g2.
40	Design change:
41	• If you’re not the leader, don’t talk to zero.
42	• Let the leader send you updates via proposals.
43
44	Move:
45	• Dgraph zero would decide that G1 should not serve P, G2 should serve it.
46	• Zero would propose that G1 is read-only for predicate P. This would propagate to the cluster.
47
48	• Zero would tell G1 to move P to G2 (Endpoint: Zero → G1)
49
50	This would trigger G1 to get latest state. Wait for it.
51	• G1 would propose this state to it’s followers.
52	• G1 after proposing would do a call to G2, and start streaming.
53	• Before G2 starts accepting, it should delete any current keys for P.
54	• It should tell Zero whether it succeeded or failed. (Endpoint: G1 → Zero)
55
56	• Zero would then propose that G2 is serving P (or G1 is, if fail above) P would RW.
57	• G1 gets this, G2 gets this.
58	• Both propagate this to their followers.
59
60	*/
61
62	// TODO: Have a event log for everything.
63	func (s *Server) rebalanceTablets() {	66✔
64	ticker := time.NewTicker(opts.rebalanceInterval)	66✔
65	for range ticker.C {	81✔
66	predicate, srcGroup, dstGroup := s.chooseTablet()	15✔
67	if len(predicate) == 0 {	30✔
68	continue	15✔
69	}
70	if err := s.movePredicate(predicate, srcGroup, dstGroup); err != nil {	×
71	glog.Errorln(err)	×
72	}	×
73	}
74	}
75
76	// MoveTablet can be used to move a tablet to a specific group.
77	// It takes in tablet and destination group as argument.
78	// It returns a *pb.Status to be used by the `/moveTablet` HTTP handler in Zero.
79	func (s Server) MoveTablet(ctx context.Context, req pb.MoveTabletRequest) (*pb.Status, error) {	1,009✔
80	if !s.Node.AmLeader() {	1,009✔
81	return &pb.Status{Code: 1, Msg: x.Error}, errNotLeader	×
82	}	×
83
84	knownGroups := s.KnownGroups()	1,009✔
85	var isKnown bool	1,009✔
86	for _, grp := range knownGroups {	2,026✔
87	if grp == req.DstGroup {	1,025✔
88	isKnown = true	8✔
89	break	8✔
90	}
91	}
92	if !isKnown {	2,010✔
93	return &pb.Status{Code: 1, Msg: x.ErrorInvalidRequest},	1,001✔
94	fmt.Errorf("group: [%d] is not a known group", req.DstGroup)	1,001✔
95	}	1,001✔
96
97	tablet := x.NamespaceAttr(req.Namespace, req.Tablet)	8✔
98	tab := s.ServingTablet(tablet)	8✔
99	if tab == nil {	8✔
100	return &pb.Status{Code: 1, Msg: x.ErrorInvalidRequest},	×
101	fmt.Errorf("namespace: %d. No tablet found for: %s", req.Namespace, req.Tablet)	×
102	}	×
103
104	srcGroup := tab.GroupId	8✔
105	if srcGroup == req.DstGroup {	13✔
106	return &pb.Status{Code: 1, Msg: x.ErrorInvalidRequest},	5✔
107	fmt.Errorf("namespace: %d. Tablet: [%s] is already being served by group: [%d]",	5✔
108	req.Namespace, req.Tablet, srcGroup)	5✔
109	}	5✔
110
111	if err := s.movePredicate(tablet, srcGroup, req.DstGroup); err != nil {	3✔
112	glog.Errorf("namespace: %d. While moving predicate %s from %d -> %d. Error: %v",	×
113	req.Namespace, req.Tablet, srcGroup, req.DstGroup, err)	×
114	return &pb.Status{Code: 1, Msg: x.Error}, err	×
115	}	×
116
117	return &pb.Status{Code: 0, Msg: fmt.Sprintf("namespace: %d. "+	3✔
118	"Predicate: [%s] moved from group [%d] to [%d]", req.Namespace, req.Tablet, srcGroup,	3✔
119	req.DstGroup)}, nil	3✔
120	}
121
122	// movePredicate is the main entry point for move predicate logic. This Zero must remain the leader
123	// for the entire duration of predicate move. If this Zero stops being the leader, the final
124	// proposal of reassigning the tablet to the destination would fail automatically.
125	func (s *Server) movePredicate(predicate string, srcGroup, dstGroup uint32) error {	3✔
126	s.moveOngoing <- struct{}{}	3✔
127	defer func() {	6✔
128	<-s.moveOngoing	3✔
129	}()	3✔
130
131	ctx, cancel := context.WithTimeout(context.Background(), predicateMoveTimeout)	3✔
132	defer cancel()	3✔
133		3✔
134	ctx, span := otrace.StartSpan(ctx, "Zero.MovePredicate")	3✔
135	defer span.End()	3✔
136		3✔
137	// Ensure that reserved predicates cannot be moved.	3✔
138	if x.IsReservedPredicate(predicate) {	3✔
139	return errors.Errorf("Unable to move reserved predicate %s", predicate)	×
140	}	×
141
142	// Ensure that I'm connected to the rest of the Zero group, and am the leader.
143	if _, err := s.latestMembershipState(ctx); err != nil {	3✔
144	return errors.Wrapf(err, "unable to reach quorum")	×
145	}	×
146	if !s.Node.AmLeader() {	3✔
147	return errors.Errorf("I am not the Zero leader")	×
148	}	×
149	tab := s.ServingTablet(predicate)	3✔
150	if tab == nil {	3✔
151	return errors.Errorf("Tablet to be moved: [%v] is not being served", predicate)	×
152	}	×
153	msg := fmt.Sprintf("Going to move predicate: [%v], size: [ondisk: %v, uncompressed: %v]"+	3✔
154	" from group %d to %d\n", predicate, humanize.IBytes(uint64(tab.OnDiskBytes)),	3✔
155	humanize.IBytes(uint64(tab.UncompressedBytes)), srcGroup, dstGroup)	3✔
156	glog.Info(msg)	3✔
157	span.Annotate([]otrace.Attribute{otrace.StringAttribute("tablet", predicate)}, msg)	3✔
158		3✔
159	// Block all commits on this predicate. Keep them blocked until we return from this function.	3✔
160	unblock := s.blockTablet(predicate)	3✔
161	defer unblock()	3✔
162		3✔
163	// Get a new timestamp, beyond which we are sure that no new txns would be committed for this	3✔
164	// predicate. Source Alpha leader must reach this timestamp before streaming the data.	3✔
165	ids, err := s.Timestamps(ctx, &pb.Num{Val: 1})	3✔
166	if err != nil \|\| ids.StartId == 0 {	3✔
167	return errors.Wrapf(err, "while leasing txn timestamp. Id: %+v", ids)	×
168	}	×
169
170	// Get connection to leader of source group.
171	pl := s.Leader(srcGroup)	3✔
172	if pl == nil {	3✔
173	return errors.Errorf("No healthy connection found to leader of group %d", srcGroup)	×
174	}	×
175	wc := pb.NewWorkerClient(pl.Get())	3✔
176	in := &pb.MovePredicatePayload{	3✔
177	Predicate: predicate,	3✔
178	SourceGid: srcGroup,	3✔
179	DestGid: dstGroup,	3✔
180	TxnTs: ids.StartId,	3✔
181	}	3✔
182	span.Annotatef(nil, "Starting move: %+v", in)	3✔
183	glog.Infof("Starting move: %+v", in)	3✔
184	if _, err := wc.MovePredicate(ctx, in); err != nil {	3✔
185	return errors.Wrapf(err, "while calling MovePredicate")	×
186	}	×
187
188	p := &pb.ZeroProposal{}	3✔
189	p.Tablet = &pb.Tablet{	3✔
190	GroupId: dstGroup,	3✔
191	Predicate: predicate,	3✔
192	OnDiskBytes: tab.OnDiskBytes,	3✔
193	UncompressedBytes: tab.UncompressedBytes,	3✔
194	Force: true,	3✔
195	MoveTs: in.TxnTs,	3✔
196	}	3✔
197	msg = fmt.Sprintf("Move at Alpha done. Now proposing: %+v", p)	3✔
198	span.Annotate(nil, msg)	3✔
199	glog.Info(msg)	3✔
200	if err := s.Node.proposeAndWait(ctx, p); err != nil {	3✔
201	return errors.Wrapf(err, "while proposing tablet reassignment. Proposal: %+v", p)	×
202	}	×
203	msg = fmt.Sprintf("Predicate move done for: [%v] from group %d to %d\n",	3✔
204	predicate, srcGroup, dstGroup)	3✔
205	glog.Info(msg)	3✔
206	span.Annotate(nil, msg)	3✔
207		3✔
208	// Now that the move has happened, we can delete the predicate from the source group. But before	3✔
209	// doing that, we should ensure the source group understands that the predicate is now being	3✔
210	// served by the destination group. For that, we pass in the expected checksum for the source	3✔
211	// group. Only once the source group membership checksum matches, would the source group delete	3✔
212	// the predicate. This ensures that it does not service any transaction after deletion of data.	3✔
213	checksums := s.groupChecksums()	3✔
214	in.ExpectedChecksum = checksums[in.SourceGid]	3✔
215	in.DestGid = 0 // Indicates deletion of predicate in the source group.	3✔
216	if _, err := wc.MovePredicate(ctx, in); err != nil {	3✔
217	msg = fmt.Sprintf("While deleting predicate [%v] in group %d. Error: %v",	×
218	in.Predicate, in.SourceGid, err)	×
219	span.Annotate(nil, msg)	×
220	glog.Warningf(msg)	×
221	} else {	3✔
222	msg = fmt.Sprintf("Deleted predicate %v in group %d", in.Predicate, in.SourceGid)	3✔
223	span.Annotate(nil, msg)	3✔
224	glog.V(1).Infof(msg)	3✔
225	}	3✔
226	return nil	3✔
227	}
228
229	func (s *Server) chooseTablet() (predicate string, srcGroup uint32, dstGroup uint32) {	15✔
230	s.RLock()	15✔
231	defer s.RUnlock()	15✔
232	if s.state == nil {	15✔
233	return	×
234	}	×
235	numGroups := len(s.state.Groups)	15✔
236	if !s.Node.AmLeader() \|\| numGroups <= 1 {	25✔
237	return	10✔
238	}	10✔
239
240	// Sort all groups by their sizes.
241	type kv struct {	5✔
242	gid uint32	5✔
243	size int64 // in bytes	5✔
244	}	5✔
245	var groups []kv	5✔
246	for k, v := range s.state.Groups {	15✔
247	space := int64(0)	10✔
248	for _, tab := range v.Tablets {	420✔
249	space += tab.OnDiskBytes	410✔
250	}	410✔
251	groups = append(groups, kv{k, space})	10✔
252	}
253	sort.Slice(groups, func(i, j int) bool {	10✔
254	return groups[i].size < groups[j].size	5✔
255	})	5✔
256
257	glog.Infof("\n\nGroups sorted by size: %+v\n\n", groups)	5✔
258	for lastGroup := numGroups - 1; lastGroup > 0; lastGroup-- {	10✔
259	srcGroup = groups[lastGroup].gid	5✔
260	dstGroup = groups[0].gid	5✔
261	sizeDiff := groups[lastGroup].size - groups[0].size	5✔
262	glog.Infof("size_diff %v\n", sizeDiff)	5✔
263	// Don't move a node unless you receive atleast one update regarding tablet size.	5✔
264	// Tablet size would have come up with leader update.	5✔
265	if !s.hasLeader(dstGroup) {	5✔
266	return	×
267	}	×
268	// We move the predicate only if the difference between size of both machines is
269	// atleast 10% of dst group.
270	if float64(sizeDiff) < 0.1*float64(groups[0].size) {	5✔
271	continue	×
272	}
273
274	// Try to find a predicate which we can move.
275	size := int64(0)	5✔
276	group := s.state.Groups[srcGroup]	5✔
277	for _, tab := range group.Tablets {	5✔
278	// Reserved predicates should always be in group 1 so do not re-balance them.	×
279	if x.IsReservedPredicate(tab.Predicate) {	×
280	continue	×
281	}
282
283	// Finds a tablet as big a possible such that on moving it dstGroup's size is
284	// less than or equal to srcGroup.
285	if tab.OnDiskBytes <= sizeDiff/2 && tab.OnDiskBytes > size {	×
286	predicate = tab.Predicate	×
287	size = tab.OnDiskBytes	×
288	}	×
289	}
290	if len(predicate) > 0 {	5✔
291	return	×
292	}	×
293	}
294	return	5✔
295	}

dgraph-io / dgraph / 5985667653

Source File Press 'n' to go to next uncovered line, 'b' for previous

Source File
Press 'n' to go to next uncovered line, 'b' for previous