summaryrefslogtreecommitdiff
path: root/core/src/test/java/org/elasticsearch/gateway/RecoveryFromGatewayIT.java
blob: dac284ee594f6c7a477a125bc095f3325c66eb39 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
/*
 * Licensed to Elasticsearch under one or more contributor
 * license agreements. See the NOTICE file distributed with
 * this work for additional information regarding copyright
 * ownership. Elasticsearch licenses this file to you 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 org.elasticsearch.gateway;

import com.carrotsearch.hppc.cursors.ObjectCursor;
import org.elasticsearch.action.admin.indices.recovery.RecoveryResponse;
import org.elasticsearch.action.admin.indices.stats.IndexStats;
import org.elasticsearch.action.admin.indices.stats.ShardStats;
import org.elasticsearch.client.Client;
import org.elasticsearch.cluster.ClusterState;
import org.elasticsearch.cluster.metadata.IndexMetaData;
import org.elasticsearch.cluster.routing.allocation.decider.EnableAllocationDecider;
import org.elasticsearch.cluster.routing.allocation.decider.ThrottlingAllocationDecider;
import org.elasticsearch.common.settings.Settings;
import org.elasticsearch.common.xcontent.XContentFactory;
import org.elasticsearch.env.Environment;
import org.elasticsearch.index.engine.Engine;
import org.elasticsearch.index.query.QueryBuilders;
import org.elasticsearch.indices.recovery.RecoveryState;
import org.elasticsearch.plugins.Plugin;
import org.elasticsearch.test.ESIntegTestCase;
import org.elasticsearch.test.ESIntegTestCase.ClusterScope;
import org.elasticsearch.test.ESIntegTestCase.Scope;
import org.elasticsearch.test.InternalTestCluster.RestartCallback;
import org.elasticsearch.test.store.MockFSDirectoryService;
import org.elasticsearch.test.store.MockFSIndexStore;

import java.util.Collection;
import java.util.HashMap;
import java.util.Map;
import java.util.stream.IntStream;

import static org.elasticsearch.cluster.metadata.IndexMetaData.SETTING_NUMBER_OF_REPLICAS;
import static org.elasticsearch.cluster.metadata.IndexMetaData.SETTING_NUMBER_OF_SHARDS;
import static org.elasticsearch.common.settings.Settings.settingsBuilder;
import static org.elasticsearch.common.xcontent.XContentFactory.jsonBuilder;
import static org.elasticsearch.index.query.QueryBuilders.matchAllQuery;
import static org.elasticsearch.index.query.QueryBuilders.termQuery;
import static org.elasticsearch.test.hamcrest.ElasticsearchAssertions.assertAcked;
import static org.elasticsearch.test.hamcrest.ElasticsearchAssertions.assertHitCount;
import static org.hamcrest.Matchers.equalTo;
import static org.hamcrest.Matchers.greaterThan;
import static org.hamcrest.Matchers.not;
import static org.hamcrest.Matchers.notNullValue;

@ClusterScope(numDataNodes = 0, scope = Scope.TEST)
public class RecoveryFromGatewayIT extends ESIntegTestCase {

    @Override
    protected Collection<Class<? extends Plugin>> nodePlugins() {
        return pluginList(MockFSIndexStore.TestPlugin.class);
    }

    public void testOneNodeRecoverFromGateway() throws Exception {

        internalCluster().startNode();

        String mapping = XContentFactory.jsonBuilder().startObject().startObject("type1")
                .startObject("properties").startObject("appAccountIds").field("type", "text").endObject().endObject()
                .endObject().endObject().string();
        assertAcked(prepareCreate("test").addMapping("type1", mapping));

        client().prepareIndex("test", "type1", "10990239").setSource(jsonBuilder().startObject()
                .startArray("appAccountIds").value(14).value(179).endArray().endObject()).execute().actionGet();
        client().prepareIndex("test", "type1", "10990473").setSource(jsonBuilder().startObject()
                .startArray("appAccountIds").value(14).endArray().endObject()).execute().actionGet();
        client().prepareIndex("test", "type1", "10990513").setSource(jsonBuilder().startObject()
                .startArray("appAccountIds").value(14).value(179).endArray().endObject()).execute().actionGet();
        client().prepareIndex("test", "type1", "10990695").setSource(jsonBuilder().startObject()
                .startArray("appAccountIds").value(14).endArray().endObject()).execute().actionGet();
        client().prepareIndex("test", "type1", "11026351").setSource(jsonBuilder().startObject()
                .startArray("appAccountIds").value(14).endArray().endObject()).execute().actionGet();

        refresh();
        assertHitCount(client().prepareSearch().setSize(0).setQuery(termQuery("appAccountIds", 179)).execute().actionGet(), 2);
        ensureYellow("test"); // wait for primary allocations here otherwise if we have a lot of shards we might have a
        // shard that is still in post recovery when we restart and the ensureYellow() below will timeout

        Map<String, long[]> primaryTerms = assertAndCapturePrimaryTerms(null);
        internalCluster().fullRestart();

        logger.info("Running Cluster Health (wait for the shards to startup)");
        ensureYellow();
        primaryTerms = assertAndCapturePrimaryTerms(primaryTerms);

        client().admin().indices().prepareRefresh().execute().actionGet();
        assertHitCount(client().prepareSearch().setSize(0).setQuery(termQuery("appAccountIds", 179)).execute().actionGet(), 2);

        internalCluster().fullRestart();

        logger.info("Running Cluster Health (wait for the shards to startup)");
        ensureYellow();
        primaryTerms = assertAndCapturePrimaryTerms(primaryTerms);

        client().admin().indices().prepareRefresh().execute().actionGet();
        assertHitCount(client().prepareSearch().setSize(0).setQuery(termQuery("appAccountIds", 179)).execute().actionGet(), 2);
    }

    private Map<String, long[]> assertAndCapturePrimaryTerms(Map<String, long[]> previousTerms) {
        if (previousTerms == null) {
            previousTerms = new HashMap<>();
        }
        final Map<String, long[]> result = new HashMap<>();
        final ClusterState state = client().admin().cluster().prepareState().get().getState();
        for (ObjectCursor<IndexMetaData> cursor : state.metaData().indices().values()) {
            final IndexMetaData indexMetaData = cursor.value;
            final String index = indexMetaData.getIndex().getName();
            final long[] previous = previousTerms.get(index);
            final long[] current = IntStream.range(0, indexMetaData.getNumberOfShards()).mapToLong(indexMetaData::primaryTerm).toArray();
            if (previous == null) {
                result.put(index, current);
            } else {
                assertThat("number of terms changed for index [" + index + "]", current.length, equalTo(previous.length));
                for (int shard = 0; shard < current.length; shard++) {
                    assertThat("primary term didn't increase for [" + index + "][" + shard + "]", current[shard], greaterThan(previous[shard]));
                }
                result.put(index, current);
            }
        }

        return result;
    }

    public void testSingleNodeNoFlush() throws Exception {
        internalCluster().startNode();

        String mapping = XContentFactory.jsonBuilder().startObject().startObject("type1")
                .startObject("properties").startObject("field").field("type", "text").endObject().startObject("num").field("type", "integer").endObject().endObject()
                .endObject().endObject().string();
        // note: default replica settings are tied to #data nodes-1 which is 0 here. We can do with 1 in this test.
        int numberOfShards = numberOfShards();
        assertAcked(prepareCreate("test").setSettings(
                SETTING_NUMBER_OF_SHARDS, numberOfShards(),
                SETTING_NUMBER_OF_REPLICAS, randomIntBetween(0, 1)
        ).addMapping("type1", mapping));

        int value1Docs;
        int value2Docs;
        boolean indexToAllShards = randomBoolean();

        if (indexToAllShards) {
            // insert enough docs so all shards will have a doc
            value1Docs = randomIntBetween(numberOfShards * 10, numberOfShards * 20);
            value2Docs = randomIntBetween(numberOfShards * 10, numberOfShards * 20);

        } else {
            // insert a two docs, some shards will not have anything
            value1Docs = 1;
            value2Docs = 1;
        }


        for (int i = 0; i < 1 + randomInt(100); i++) {
            for (int id = 0; id < Math.max(value1Docs, value2Docs); id++) {
                if (id < value1Docs) {
                    index("test", "type1", "1_" + id,
                            jsonBuilder().startObject().field("field", "value1").startArray("num").value(14).value(179).endArray().endObject()
                    );
                }
                if (id < value2Docs) {
                    index("test", "type1", "2_" + id,
                            jsonBuilder().startObject().field("field", "value2").startArray("num").value(14).endArray().endObject()
                    );
                }
            }

        }

        refresh();

        for (int i = 0; i <= randomInt(10); i++) {
            assertHitCount(client().prepareSearch().setSize(0).setQuery(matchAllQuery()).get(), value1Docs + value2Docs);
            assertHitCount(client().prepareSearch().setSize(0).setQuery(termQuery("field", "value1")).get(), value1Docs);
            assertHitCount(client().prepareSearch().setSize(0).setQuery(termQuery("field", "value2")).get(), value2Docs);
            assertHitCount(client().prepareSearch().setSize(0).setQuery(termQuery("num", 179)).get(), value1Docs);
        }
        if (!indexToAllShards) {
            // we have to verify primaries are started for them to be restored
            logger.info("Ensure all primaries have been started");
            ensureYellow();
        }

        Map<String, long[]> primaryTerms = assertAndCapturePrimaryTerms(null);

        internalCluster().fullRestart();

        logger.info("Running Cluster Health (wait for the shards to startup)");
        ensureYellow();
        primaryTerms = assertAndCapturePrimaryTerms(primaryTerms);

        for (int i = 0; i <= randomInt(10); i++) {
            assertHitCount(client().prepareSearch().setSize(0).setQuery(matchAllQuery()).get(), value1Docs + value2Docs);
            assertHitCount(client().prepareSearch().setSize(0).setQuery(termQuery("field", "value1")).get(), value1Docs);
            assertHitCount(client().prepareSearch().setSize(0).setQuery(termQuery("field", "value2")).get(), value2Docs);
            assertHitCount(client().prepareSearch().setSize(0).setQuery(termQuery("num", 179)).get(), value1Docs);
        }

        internalCluster().fullRestart();


        logger.info("Running Cluster Health (wait for the shards to startup)");
        ensureYellow();
        primaryTerms = assertAndCapturePrimaryTerms(primaryTerms);

        for (int i = 0; i <= randomInt(10); i++) {
            assertHitCount(client().prepareSearch().setSize(0).setQuery(matchAllQuery()).get(), value1Docs + value2Docs);
            assertHitCount(client().prepareSearch().setSize(0).setQuery(termQuery("field", "value1")).get(), value1Docs);
            assertHitCount(client().prepareSearch().setSize(0).setQuery(termQuery("field", "value2")).get(), value2Docs);
            assertHitCount(client().prepareSearch().setSize(0).setQuery(termQuery("num", 179)).get(), value1Docs);
        }
    }

    public void testSingleNodeWithFlush() throws Exception {
        internalCluster().startNode();
        client().prepareIndex("test", "type1", "1").setSource(jsonBuilder().startObject().field("field", "value1").endObject()).execute().actionGet();
        flush();
        client().prepareIndex("test", "type1", "2").setSource(jsonBuilder().startObject().field("field", "value2").endObject()).execute().actionGet();
        refresh();

        assertHitCount(client().prepareSearch().setSize(0).setQuery(matchAllQuery()).execute().actionGet(), 2);

        ensureYellow("test"); // wait for primary allocations here otherwise if we have a lot of shards we might have a
        // shard that is still in post recovery when we restart and the ensureYellow() below will timeout

        Map<String, long[]> primaryTerms = assertAndCapturePrimaryTerms(null);

        internalCluster().fullRestart();

        logger.info("Running Cluster Health (wait for the shards to startup)");
        ensureYellow();
        primaryTerms = assertAndCapturePrimaryTerms(primaryTerms);

        for (int i = 0; i < 10; i++) {
            assertHitCount(client().prepareSearch().setSize(0).setQuery(matchAllQuery()).execute().actionGet(), 2);
        }

        internalCluster().fullRestart();

        logger.info("Running Cluster Health (wait for the shards to startup)");
        ensureYellow();
        primaryTerms = assertAndCapturePrimaryTerms(primaryTerms);

        for (int i = 0; i < 10; i++) {
            assertHitCount(client().prepareSearch().setSize(0).setQuery(matchAllQuery()).execute().actionGet(), 2);
        }
    }

    public void testTwoNodeFirstNodeCleared() throws Exception {
        final String firstNode = internalCluster().startNode();
        internalCluster().startNode();

        client().prepareIndex("test", "type1", "1").setSource(jsonBuilder().startObject().field("field", "value1").endObject()).execute().actionGet();
        flush();
        client().prepareIndex("test", "type1", "2").setSource(jsonBuilder().startObject().field("field", "value2").endObject()).execute().actionGet();
        refresh();

        logger.info("Running Cluster Health (wait for the shards to startup)");
        ensureGreen();

        for (int i = 0; i < 10; i++) {
            assertHitCount(client().prepareSearch().setSize(0).setQuery(matchAllQuery()).execute().actionGet(), 2);
        }

        Map<String, long[]> primaryTerms = assertAndCapturePrimaryTerms(null);

        internalCluster().fullRestart(new RestartCallback() {
            @Override
            public Settings onNodeStopped(String nodeName) throws Exception {
                return settingsBuilder().put("gateway.recover_after_nodes", 2).build();
            }

            @Override
            public boolean clearData(String nodeName) {
                return firstNode.equals(nodeName);
            }

        });

        logger.info("Running Cluster Health (wait for the shards to startup)");
        ensureGreen();
        primaryTerms = assertAndCapturePrimaryTerms(primaryTerms);

        for (int i = 0; i < 10; i++) {
            assertHitCount(client().prepareSearch().setSize(0).setQuery(matchAllQuery()).execute().actionGet(), 2);
        }
    }

    public void testLatestVersionLoaded() throws Exception {
        // clean two nodes
        internalCluster().startNodesAsync(2, settingsBuilder().put("gateway.recover_after_nodes", 2).build()).get();

        client().prepareIndex("test", "type1", "1").setSource(jsonBuilder().startObject().field("field", "value1").endObject()).execute().actionGet();
        client().admin().indices().prepareFlush().execute().actionGet();
        client().prepareIndex("test", "type1", "2").setSource(jsonBuilder().startObject().field("field", "value2").endObject()).execute().actionGet();
        client().admin().indices().prepareRefresh().execute().actionGet();

        logger.info("--> running cluster_health (wait for the shards to startup)");
        ensureGreen();

        for (int i = 0; i < 10; i++) {
            assertHitCount(client().prepareSearch().setSize(0).setQuery(matchAllQuery()).execute().actionGet(), 2);
        }

        String metaDataUuid = client().admin().cluster().prepareState().execute().get().getState().getMetaData().clusterUUID();
        assertThat(metaDataUuid, not(equalTo("_na_")));

        Map<String, long[]> primaryTerms = assertAndCapturePrimaryTerms(null);

        logger.info("--> closing first node, and indexing more data to the second node");
        internalCluster().fullRestart(new RestartCallback() {

            @Override
            public void doAfterNodes(int numNodes, Client client) throws Exception {
                if (numNodes == 1) {
                    logger.info("--> one node is closed - start indexing data into the second one");
                    client.prepareIndex("test", "type1", "3").setSource(jsonBuilder().startObject().field("field", "value3").endObject()).execute().actionGet();
                    // TODO: remove once refresh doesn't fail immediately if there a master block:
                    // https://github.com/elastic/elasticsearch/issues/9997
                    client.admin().cluster().prepareHealth("test").setWaitForYellowStatus().get();
                    client.admin().indices().prepareRefresh().execute().actionGet();

                    for (int i = 0; i < 10; i++) {
                        assertHitCount(client.prepareSearch().setSize(0).setQuery(matchAllQuery()).execute().actionGet(), 3);
                    }

                    logger.info("--> add some metadata, additional type and template");
                    client.admin().indices().preparePutMapping("test").setType("type2")
                            .setSource(jsonBuilder().startObject().startObject("type2").endObject().endObject())
                            .execute().actionGet();
                    client.admin().indices().preparePutTemplate("template_1")
                            .setTemplate("te*")
                            .setOrder(0)
                            .addMapping("type1", XContentFactory.jsonBuilder().startObject().startObject("type1").startObject("properties")
                                    .startObject("field1").field("type", "text").field("store", true).endObject()
                                    .startObject("field2").field("type", "keyword").field("store", true).endObject()
                                    .endObject().endObject().endObject())
                            .execute().actionGet();
                    client.admin().indices().prepareAliases().addAlias("test", "test_alias", QueryBuilders.termQuery("field", "value")).execute().actionGet();
                    logger.info("--> starting two nodes back, verifying we got the latest version");
                }

            }

        });

        logger.info("--> running cluster_health (wait for the shards to startup)");
        ensureGreen();
        primaryTerms = assertAndCapturePrimaryTerms(primaryTerms);

        assertThat(client().admin().cluster().prepareState().execute().get().getState().getMetaData().clusterUUID(), equalTo(metaDataUuid));

        for (int i = 0; i < 10; i++) {
            assertHitCount(client().prepareSearch().setSize(0).setQuery(matchAllQuery()).execute().actionGet(), 3);
        }

        ClusterState state = client().admin().cluster().prepareState().execute().actionGet().getState();
        assertThat(state.metaData().index("test").mapping("type2"), notNullValue());
        assertThat(state.metaData().templates().get("template_1").template(), equalTo("te*"));
        assertThat(state.metaData().index("test").getAliases().get("test_alias"), notNullValue());
        assertThat(state.metaData().index("test").getAliases().get("test_alias").filter(), notNullValue());
    }

    public void testReusePeerRecovery() throws Exception {
        final Settings settings = settingsBuilder()
                .put(MockFSIndexStore.INDEX_CHECK_INDEX_ON_CLOSE_SETTING.getKey(), false)
                .put("gateway.recover_after_nodes", 4)
                .put(ThrottlingAllocationDecider.CLUSTER_ROUTING_ALLOCATION_NODE_CONCURRENT_INCOMING_RECOVERIES_SETTING.getKey(), 4)
                .put(ThrottlingAllocationDecider.CLUSTER_ROUTING_ALLOCATION_NODE_CONCURRENT_OUTGOING_RECOVERIES_SETTING.getKey(), 4)
                .put(MockFSDirectoryService.CRASH_INDEX_SETTING.getKey(), false).build();

        internalCluster().startNodesAsync(4, settings).get();
        // prevent any rebalance actions during the peer recovery
        // if we run into a relocation the reuse count will be 0 and this fails the test. We are testing here if
        // we reuse the files on disk after full restarts for replicas.
        assertAcked(prepareCreate("test").setSettings(Settings.builder()
                .put(indexSettings())
                .put(EnableAllocationDecider.INDEX_ROUTING_REBALANCE_ENABLE_SETTING.getKey(), EnableAllocationDecider.Rebalance.NONE)));
        ensureGreen();
        logger.info("--> indexing docs");
        for (int i = 0; i < 1000; i++) {
            client().prepareIndex("test", "type").setSource("field", "value").execute().actionGet();
            if ((i % 200) == 0) {
                client().admin().indices().prepareFlush().execute().actionGet();
            }
        }
        if (randomBoolean()) {
            client().admin().indices().prepareFlush().execute().actionGet();
        }
        logger.info("Running Cluster Health");
        ensureGreen();
        client().admin().indices().prepareForceMerge("test").setMaxNumSegments(100).get(); // just wait for merges
        client().admin().indices().prepareFlush().setWaitIfOngoing(true).setForce(true).get();

        boolean useSyncIds = randomBoolean();
        if (useSyncIds == false) {
            logger.info("--> disabling allocation while the cluster is shut down");

            // Disable allocations while we are closing nodes
            client().admin().cluster().prepareUpdateSettings()
                    .setTransientSettings(settingsBuilder()
                            .put(EnableAllocationDecider.CLUSTER_ROUTING_ALLOCATION_ENABLE_SETTING.getKey(), EnableAllocationDecider.Allocation.NONE))
                    .get();
            logger.info("--> full cluster restart");
            internalCluster().fullRestart();

            logger.info("--> waiting for cluster to return to green after first shutdown");
            ensureGreen();
        } else {
            logger.info("--> trying to sync flush");
            assertEquals(client().admin().indices().prepareSyncedFlush("test").get().failedShards(), 0);
            assertSyncIdsNotNull();
        }

        logger.info("--> disabling allocation while the cluster is shut down{}", useSyncIds ? "" : " a second time");
        // Disable allocations while we are closing nodes
        client().admin().cluster().prepareUpdateSettings()
                .setTransientSettings(settingsBuilder()
                        .put(EnableAllocationDecider.CLUSTER_ROUTING_ALLOCATION_ENABLE_SETTING.getKey(), EnableAllocationDecider.Allocation.NONE))
                .get();

        Map<String, long[]> primaryTerms = assertAndCapturePrimaryTerms(null);

        logger.info("--> full cluster restart");
        internalCluster().fullRestart();

        logger.info("--> waiting for cluster to return to green after {}shutdown", useSyncIds ? "" : "second ");
        ensureGreen();
        primaryTerms = assertAndCapturePrimaryTerms(primaryTerms);

        if (useSyncIds) {
            assertSyncIdsNotNull();
        }
        RecoveryResponse recoveryResponse = client().admin().indices().prepareRecoveries("test").get();
        for (RecoveryState recoveryState : recoveryResponse.shardRecoveryStates().get("test")) {
            long recovered = 0;
            for (RecoveryState.File file : recoveryState.getIndex().fileDetails()) {
                if (file.name().startsWith("segments")) {
                    recovered += file.length();
                }
            }
            if (!recoveryState.getPrimary() && (useSyncIds == false)) {
                logger.info("--> replica shard {} recovered from {} to {}, recovered {}, reuse {}",
                        recoveryState.getShardId().getId(), recoveryState.getSourceNode().name(), recoveryState.getTargetNode().name(),
                        recoveryState.getIndex().recoveredBytes(), recoveryState.getIndex().reusedBytes());
                assertThat("no bytes should be recovered", recoveryState.getIndex().recoveredBytes(), equalTo(recovered));
                assertThat("data should have been reused", recoveryState.getIndex().reusedBytes(), greaterThan(0L));
                // we have to recover the segments file since we commit the translog ID on engine startup
                assertThat("all bytes should be reused except of the segments file", recoveryState.getIndex().reusedBytes(), equalTo(recoveryState.getIndex().totalBytes() - recovered));
                assertThat("no files should be recovered except of the segments file", recoveryState.getIndex().recoveredFileCount(), equalTo(1));
                assertThat("all files should be reused except of the segments file", recoveryState.getIndex().reusedFileCount(), equalTo(recoveryState.getIndex().totalFileCount() - 1));
                assertThat("> 0 files should be reused", recoveryState.getIndex().reusedFileCount(), greaterThan(0));
            } else {
                if (useSyncIds && !recoveryState.getPrimary()) {
                    logger.info("--> replica shard {} recovered from {} to {} using sync id, recovered {}, reuse {}",
                            recoveryState.getShardId().getId(), recoveryState.getSourceNode().name(), recoveryState.getTargetNode().name(),
                            recoveryState.getIndex().recoveredBytes(), recoveryState.getIndex().reusedBytes());
                }
                assertThat(recoveryState.getIndex().recoveredBytes(), equalTo(0L));
                assertThat(recoveryState.getIndex().reusedBytes(), equalTo(recoveryState.getIndex().totalBytes()));
                assertThat(recoveryState.getIndex().recoveredFileCount(), equalTo(0));
                assertThat(recoveryState.getIndex().reusedFileCount(), equalTo(recoveryState.getIndex().totalFileCount()));
            }
        }
    }

    public void assertSyncIdsNotNull() {
        IndexStats indexStats = client().admin().indices().prepareStats("test").get().getIndex("test");
        for (ShardStats shardStats : indexStats.getShards()) {
            assertNotNull(shardStats.getCommitStats().getUserData().get(Engine.SYNC_COMMIT_ID));
        }
    }

    public void testRecoveryDifferentNodeOrderStartup() throws Exception {
        // we need different data paths so we make sure we start the second node fresh

        final String node_1 = internalCluster().startNode(settingsBuilder().put(Environment.PATH_DATA_SETTING.getKey(), createTempDir()).build());

        client().prepareIndex("test", "type1", "1").setSource("field", "value").execute().actionGet();

        internalCluster().startNode(settingsBuilder().put(Environment.PATH_DATA_SETTING.getKey(), createTempDir()).build());

        ensureGreen();
        Map<String, long[]> primaryTerms = assertAndCapturePrimaryTerms(null);


        internalCluster().fullRestart(new RestartCallback() {

            @Override
            public boolean doRestart(String nodeName) {
                return !node_1.equals(nodeName);
            }
        });

        ensureYellow();
        primaryTerms = assertAndCapturePrimaryTerms(primaryTerms);

        assertThat(client().admin().indices().prepareExists("test").execute().actionGet().isExists(), equalTo(true));
        assertHitCount(client().prepareSearch("test").setSize(0).setQuery(QueryBuilders.matchAllQuery()).execute().actionGet(), 1);
    }

}