honeymoose
/
OpenSearch
mirror of https://github.com/honeymoose/OpenSearch.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

Wait for mapping in testReadRequestsReturnLatestMappingVersion (#37886) 

If the index request is executed before the mapping update is applied on
the IndexShard, the index request will perform a dynamic mapping update.
This mapping update will be timeout (i.e, ProcessClusterEventTimeoutException)
because the latch is not open. This leads to the failure of the index
request and the test. This commit makes sure the mapping is ready
before we execute the index request.

Closes #37807
This commit is contained in:
Nhat Nguyen 2019-01-28 15:25:56 -05:00 committed by GitHub
parent 6325e55dbf
commit 557fcf915e
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
1 changed files with 30 additions and 19 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

49
x-pack/plugin/ccr/src/test/java/org/elasticsearch/xpack/ccr/FollowerFailOverIT.java
View File

@ -22,6 +22,7 @@ import org.elasticsearch.common.unit.TimeValue;
import org.elasticsearch.common.xcontent.XContentType;
import org.elasticsearch.common.xcontent.support.XContentMapValues;
import org.elasticsearch.index.IndexSettings;
import org.elasticsearch.index.mapper.DocumentMapper;
import org.elasticsearch.index.shard.IndexShard;
import org.elasticsearch.index.shard.ShardId;
import org.elasticsearch.indices.IndicesService;
@ -230,8 +231,7 @@ public class FollowerFailOverIT extends CcrIntegTestCase {
pauseFollow("follower-index");
}
@AwaitsFix(bugUrl = "https://github.com/elastic/elasticsearch/issues/37807")
public void testReadRequestsReturnsLatestMappingVersion() throws Exception {
public void testReadRequestsReturnLatestMappingVersion() throws Exception {
InternalTestCluster leaderCluster = getLeaderCluster();
Settings nodeAttributes = Settings.builder().put("node.attr.box", "large").build();
String dataNode = leaderCluster.startDataOnlyNode(nodeAttributes);
@ -244,6 +244,9 @@ public class FollowerFailOverIT extends CcrIntegTestCase {
.put("index.routing.allocation.require.box", "large"))
.get()
);
getFollowerCluster().startDataOnlyNode(nodeAttributes);
followerClient().execute(PutFollowAction.INSTANCE, putFollow("leader-index", "follower-index")).get();
ensureFollowerGreen("follower-index");
ClusterService clusterService = leaderCluster.clusterService(dataNode);
ShardId shardId = clusterService.state().routingTable().index("leader-index").shard(0).shardId();
IndicesService indicesService = leaderCluster.getInstance(IndicesService.class, dataNode);
@ -265,22 +268,30 @@ public class FollowerFailOverIT extends CcrIntegTestCase {
});
leaderCluster.client().admin().indices().preparePutMapping().setType("doc")
.setSource("balance", "type=long").setTimeout(TimeValue.ZERO).get();
IndexResponse indexResp = leaderCluster.client(dataNode).prepareIndex("leader-index", "doc", "1")
.setSource("{\"balance\": 100}", XContentType.JSON).setTimeout(TimeValue.ZERO).get();
assertThat(indexResp.getResult(), equalTo(DocWriteResponse.Result.CREATED));
assertThat(indexShard.getGlobalCheckpoint(), equalTo(0L));
getFollowerCluster().startDataOnlyNode(nodeAttributes);
followerClient().execute(PutFollowAction.INSTANCE, putFollow("leader-index", "follower-index")).get();
ensureFollowerGreen("follower-index");
// Make sure at least one read-request which requires mapping sync is completed.
assertBusy(() -> {
CcrClient ccrClient = new CcrClient(followerClient());
FollowStatsAction.StatsResponses responses = ccrClient.followStats(new FollowStatsAction.StatsRequest()).actionGet();
long bytesRead = responses.getStatsResponses().stream().mapToLong(r -> r.status().bytesRead()).sum();
assertThat(bytesRead, Matchers.greaterThan(0L));
}, 60, TimeUnit.SECONDS);
latch.countDown();
assertIndexFullyReplicatedToFollower("leader-index", "follower-index");
pauseFollow("follower-index");
try {
// Make sure the mapping is ready on the shard before we execute the index request; otherwise the index request
// will perform a dynamic mapping update which however will be blocked because the latch is remained closed.
assertBusy(() -> {
DocumentMapper mapper = indexShard.mapperService().documentMapper("doc");
assertNotNull(mapper);
assertNotNull(mapper.mappers().getMapper("balance"));
});
IndexResponse indexResp = leaderCluster.client().prepareIndex("leader-index", "doc", "1")
.setSource("{\"balance\": 100}", XContentType.JSON).setTimeout(TimeValue.ZERO).get();
assertThat(indexResp.getResult(), equalTo(DocWriteResponse.Result.CREATED));
assertThat(indexShard.getGlobalCheckpoint(), equalTo(0L));
// Make sure at least one read-request which requires mapping sync is completed.
assertBusy(() -> {
CcrClient ccrClient = new CcrClient(followerClient());
FollowStatsAction.StatsResponses responses = ccrClient.followStats(new FollowStatsAction.StatsRequest()).actionGet();
long bytesRead = responses.getStatsResponses().stream().mapToLong(r -> r.status().bytesRead()).sum();
assertThat(bytesRead, Matchers.greaterThan(0L));
}, 60, TimeUnit.SECONDS);
latch.countDown();
assertIndexFullyReplicatedToFollower("leader-index", "follower-index");
} finally {
latch.countDown(); // no effect if latch was counted down - this makes sure teardown can make progress.
pauseFollow("follower-index");
}
}
}