Skip to content

Commit 9cd62ab

Browse files
tlrxtlrx
committed
TransportResyncReplicationAction should not honour blocks (#35795)
After #35332 has been merged, we noticed some test failures like #35597 in which one or more replica shards failed to be promoted as primaries because the primary replica re-synchronization never succeed. After some digging it appeared that the execution of the resync action was blocked because of the presence of a global cluster block in the cluster state (in this case, the "no master" block), making the resync action to fail when executed on the primary. Until #35332 such failures never happened because the TransportResyncReplicationAction is skipping the reroute phase, the only place where blocks were checked. Now with #35332 blocks are checked during reroute and also during the execution of the transport replication action on the primary. After some internal discussion, we decided that the TransportResyncReplicationAction should never be blocked. This action is part of the replica to primary promotion and makes sure that replicas are in sync and should not be blocked when the cluster state has no master or when the index is read only. This commit changes the TransportResyncReplicationAction to make obvious that it does not honor blocks. It also adds a simple test that fails if the resync action is blocked during the primary action execution. Closes #35597
1 parent f95edcb commit 9cd62ab

File tree

2 files changed

+174
-0
lines changed

2 files changed

+174
-0
lines changed

‎server/src/main/java/org/elasticsearch/action/resync/TransportResyncReplicationAction.java‎

Lines changed: 13 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -27,6 +27,7 @@
2727
import org.elasticsearch.action.support.replication.TransportReplicationAction;
2828
import org.elasticsearch.action.support.replication.TransportWriteAction;
2929
import org.elasticsearch.cluster.action.shard.ShardStateAction;
30+
import org.elasticsearch.cluster.block.ClusterBlockLevel;
3031
import org.elasticsearch.cluster.metadata.IndexNameExpressionResolver;
3132
import org.elasticsearch.cluster.node.DiscoveryNode;
3233
import org.elasticsearch.cluster.routing.ShardRouting;
@@ -102,6 +103,18 @@ protected void sendReplicaRequest(
102103
}
103104
}
104105

106+
@Override
107+
protected ClusterBlockLevel globalBlockLevel() {
108+
// resync should never be blocked because it's an internal action
109+
return null;
110+
}
111+
112+
@Override
113+
protected ClusterBlockLevel indexBlockLevel() {
114+
// resync should never be blocked because it's an internal action
115+
return null;
116+
}
117+
105118
@Override
106119
protected WritePrimaryResult<ResyncReplicationRequest, ResyncReplicationResponse> shardOperationOnPrimary(
107120
ResyncReplicationRequest request, IndexShard primary) throws Exception {

‎server/src/test/java/org/elasticsearch/action/resync/TransportResyncReplicationActionTests.java‎

Lines changed: 161 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,161 @@
1+
/*
2+
* Licensed to Elasticsearch under one or more contributor
3+
* license agreements. See the NOTICE file distributed with
4+
* this work for additional information regarding copyright
5+
* ownership. Elasticsearch licenses this file to you under
6+
* the Apache License, Version 2.0 (the "License"); you may
7+
* not use this file except in compliance with the License.
8+
* You may obtain a copy of the License at
9+
*
10+
* http://www.apache.org/licenses/LICENSE-2.0
11+
*
12+
* Unless required by applicable law or agreed to in writing,
13+
* software distributed under the License is distributed on an
14+
* "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
15+
* KIND, either express or implied. See the License for the
16+
* specific language governing permissions and limitations
17+
* under the License.
18+
*/
19+
package org.elasticsearch.action.resync;
20+
21+
import org.elasticsearch.action.ActionListener;
22+
import org.elasticsearch.action.support.ActionFilters;
23+
import org.elasticsearch.action.support.PlainActionFuture;
24+
import org.elasticsearch.cluster.ClusterState;
25+
import org.elasticsearch.cluster.action.shard.ShardStateAction;
26+
import org.elasticsearch.cluster.block.ClusterBlocks;
27+
import org.elasticsearch.cluster.metadata.IndexMetaData;
28+
import org.elasticsearch.cluster.metadata.IndexNameExpressionResolver;
29+
import org.elasticsearch.cluster.routing.IndexShardRoutingTable;
30+
import org.elasticsearch.cluster.routing.ShardRouting;
31+
import org.elasticsearch.cluster.routing.ShardRoutingState;
32+
import org.elasticsearch.cluster.service.ClusterService;
33+
import org.elasticsearch.common.io.stream.NamedWriteableRegistry;
34+
import org.elasticsearch.common.lease.Releasable;
35+
import org.elasticsearch.common.network.NetworkService;
36+
import org.elasticsearch.common.settings.Settings;
37+
import org.elasticsearch.common.util.BigArrays;
38+
import org.elasticsearch.discovery.DiscoverySettings;
39+
import org.elasticsearch.index.Index;
40+
import org.elasticsearch.index.IndexService;
41+
import org.elasticsearch.index.shard.IndexShard;
42+
import org.elasticsearch.index.shard.ReplicationGroup;
43+
import org.elasticsearch.index.shard.ShardId;
44+
import org.elasticsearch.index.translog.Translog;
45+
import org.elasticsearch.indices.IndicesService;
46+
import org.elasticsearch.indices.breaker.NoneCircuitBreakerService;
47+
import org.elasticsearch.tasks.Task;
48+
import org.elasticsearch.test.ESTestCase;
49+
import org.elasticsearch.test.transport.MockTransportService;
50+
import org.elasticsearch.threadpool.TestThreadPool;
51+
import org.elasticsearch.threadpool.ThreadPool;
52+
import org.elasticsearch.transport.MockTcpTransport;
53+
import org.junit.AfterClass;
54+
import org.junit.BeforeClass;
55+
56+
import java.nio.charset.Charset;
57+
import java.util.Collections;
58+
import java.util.HashSet;
59+
import java.util.concurrent.TimeUnit;
60+
61+
import static java.util.Collections.emptyList;
62+
import static org.elasticsearch.action.support.replication.ClusterStateCreationUtils.state;
63+
import static org.elasticsearch.test.ClusterServiceUtils.createClusterService;
64+
import static org.elasticsearch.test.ClusterServiceUtils.setState;
65+
import static org.elasticsearch.transport.TransportService.NOOP_TRANSPORT_INTERCEPTOR;
66+
import static org.hamcrest.Matchers.equalTo;
67+
import static org.hamcrest.Matchers.is;
68+
import static org.hamcrest.Matchers.nullValue;
69+
import static org.mockito.Matchers.any;
70+
import static org.mockito.Matchers.anyObject;
71+
import static org.mockito.Matchers.anyString;
72+
import static org.mockito.Matchers.eq;
73+
import static org.mockito.Mockito.doAnswer;
74+
import static org.mockito.Mockito.mock;
75+
import static org.mockito.Mockito.when;
76+
77+
public class TransportResyncReplicationActionTests extends ESTestCase {
78+
79+
private static ThreadPool threadPool;
80+
81+
@BeforeClass
82+
public static void beforeClass() {
83+
threadPool = new TestThreadPool("ShardReplicationTests");
84+
}
85+
86+
@AfterClass
87+
public static void afterClass() {
88+
ThreadPool.terminate(threadPool, 30, TimeUnit.SECONDS);
89+
threadPool = null;
90+
}
91+
92+
public void testResyncDoesNotBlockOnPrimaryAction() throws Exception {
93+
try (ClusterService clusterService = createClusterService(threadPool)) {
94+
final String indexName = randomAlphaOfLength(5);
95+
setState(clusterService, state(indexName, true, ShardRoutingState.STARTED));
96+
97+
setState(clusterService,
98+
ClusterState.builder(clusterService.state()).blocks(ClusterBlocks.builder()
99+
.addGlobalBlock(DiscoverySettings.NO_MASTER_BLOCK_ALL)
100+
.addIndexBlock(indexName, IndexMetaData.INDEX_WRITE_BLOCK)));
101+
102+
try (MockTcpTransport transport = new MockTcpTransport(Settings.EMPTY, threadPool, BigArrays.NON_RECYCLING_INSTANCE,
103+
new NoneCircuitBreakerService(), new NamedWriteableRegistry(emptyList()), new NetworkService(emptyList()))) {
104+
105+
final MockTransportService transportService = new MockTransportService(Settings.EMPTY, transport, threadPool,
106+
NOOP_TRANSPORT_INTERCEPTOR, x -> clusterService.localNode(), null, Collections.emptySet());
107+
transportService.start();
108+
transportService.acceptIncomingRequests();
109+
final ShardStateAction shardStateAction = new ShardStateAction(clusterService, transportService, null, null, threadPool);
110+
111+
final IndexMetaData indexMetaData = clusterService.state().metaData().index(indexName);
112+
final Index index = indexMetaData.getIndex();
113+
final ShardId shardId = new ShardId(index, 0);
114+
final IndexShardRoutingTable shardRoutingTable = clusterService.state().routingTable().shardRoutingTable(shardId);
115+
final ShardRouting primaryShardRouting = clusterService.state().routingTable().shardRoutingTable(shardId).primaryShard();
116+
final String allocationId = primaryShardRouting.allocationId().getId();
117+
final long primaryTerm = indexMetaData.primaryTerm(shardId.id());
118+
119+
final IndexShard indexShard = mock(IndexShard.class);
120+
when(indexShard.shardId()).thenReturn(shardId);
121+
when(indexShard.routingEntry()).thenReturn(primaryShardRouting);
122+
when(indexShard.getPendingPrimaryTerm()).thenReturn(primaryTerm);
123+
doAnswer(invocation -> {
124+
ActionListener<Releasable> callback = (ActionListener<Releasable>) invocation.getArguments()[0];
125+
callback.onResponse(() -> logger.trace("released"));
126+
return null;
127+
}).when(indexShard).acquirePrimaryOperationPermit(any(ActionListener.class), anyString(), anyObject());
128+
when(indexShard.getReplicationGroup()).thenReturn(
129+
new ReplicationGroup(shardRoutingTable,
130+
clusterService.state().metaData().index(index).inSyncAllocationIds(shardId.id()),
131+
shardRoutingTable.getAllAllocationIds()));
132+
133+
final IndexService indexService = mock(IndexService.class);
134+
when(indexService.getShard(eq(shardId.id()))).thenReturn(indexShard);
135+
136+
final IndicesService indexServices = mock(IndicesService.class);
137+
when(indexServices.indexServiceSafe(eq(index))).thenReturn(indexService);
138+
139+
final IndexNameExpressionResolver resolver = new IndexNameExpressionResolver(Settings.EMPTY);
140+
final TransportResyncReplicationAction action = new TransportResyncReplicationAction(Settings.EMPTY, transportService,
141+
clusterService, indexServices, threadPool, shardStateAction, new ActionFilters(new HashSet<>()), resolver);
142+
143+
assertThat(action.globalBlockLevel(), nullValue());
144+
assertThat(action.indexBlockLevel(), nullValue());
145+
146+
final Task task = mock(Task.class);
147+
when(task.getId()).thenReturn(randomNonNegativeLong());
148+
149+
final byte[] bytes = "{}".getBytes(Charset.forName("UTF-8"));
150+
final ResyncReplicationRequest request = new ResyncReplicationRequest(shardId, 42L, 100,
151+
new Translog.Operation[]{new Translog.Index("type", "id", 0, primaryTerm, bytes)});
152+
153+
final PlainActionFuture<ResyncReplicationResponse> listener = new PlainActionFuture<>();
154+
action.sync(request, task, allocationId, primaryTerm, listener);
155+
156+
assertThat(listener.get().getShardInfo().getFailed(), equalTo(0));
157+
assertThat(listener.isDone(), is(true));
158+
}
159+
}
160+
}
161+
}

0 commit comments

Comments
 (0)