Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

[Backport 2.15] [Remote Store] Fix sleep time bug during remote store sync #14266

Closed
wants to merge 1 commit into from
Closed

[Backport 2.15] [Remote Store] Fix sleep time bug during remote store sync #14266

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
32 changes: 29 additions & 3 deletions ...er/src/internalClusterTest/java/org/opensearch/remotemigration/MigrationBaseTestCase.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -9,18 +9,24 @@
package org.opensearch.remotemigration;

import org.opensearch.action.DocWriteResponse;
import org.opensearch.action.admin.cluster.health.ClusterHealthRequest;
import org.opensearch.action.admin.cluster.health.ClusterHealthResponse;
import org.opensearch.action.admin.cluster.repositories.get.GetRepositoriesRequest;
import org.opensearch.action.admin.cluster.repositories.get.GetRepositoriesResponse;
import org.opensearch.action.bulk.BulkRequest;
import org.opensearch.action.bulk.BulkResponse;
import org.opensearch.action.delete.DeleteResponse;
import org.opensearch.action.index.IndexRequest;
import org.opensearch.action.index.IndexResponse;
import org.opensearch.client.Requests;
import org.opensearch.cluster.ClusterState;
import org.opensearch.cluster.health.ClusterHealthStatus;
import org.opensearch.cluster.metadata.RepositoryMetadata;
import org.opensearch.cluster.routing.RoutingNode;
import org.opensearch.common.Priority;
import org.opensearch.common.UUIDs;
import org.opensearch.common.settings.Settings;
import org.opensearch.common.unit.TimeValue;
import org.opensearch.repositories.fs.ReloadableFsRepository;
import org.opensearch.test.OpenSearchIntegTestCase;
import org.junit.Before;
Expand All @@ -39,6 +45,7 @@
import static org.opensearch.node.remotestore.RemoteStoreNodeService.REMOTE_STORE_COMPATIBILITY_MODE_SETTING;
import static org.opensearch.repositories.fs.ReloadableFsRepository.REPOSITORIES_FAILRATE_SETTING;
import static org.opensearch.test.hamcrest.OpenSearchAssertions.assertAcked;
import static org.hamcrest.Matchers.equalTo;

public class MigrationBaseTestCase extends OpenSearchIntegTestCase {
protected static final String REPOSITORY_NAME = "test-remote-store-repo";
Expand Down Expand Up @@ -114,6 +121,10 @@ public void initDocRepToRemoteMigration() {
);
}

public ClusterHealthStatus ensureGreen(String... indices) {
return ensureGreen(TimeValue.timeValueSeconds(60), indices);
}

public BulkResponse indexBulk(String indexName, int numDocs) {
BulkRequest bulkRequest = new BulkRequest();
for (int i = 0; i < numDocs; i++) {
Expand Down Expand Up @@ -181,14 +192,12 @@ private Thread getIndexingThread() {
long currentDocCount = indexedDocs.incrementAndGet();
if (currentDocCount > 0 && currentDocCount % refreshFrequency == 0) {
if (rarely()) {
logger.info("--> [iteration {}] flushing index", currentDocCount);
client().admin().indices().prepareFlush(indexName).get();
logger.info("Completed ingestion of {} docs. Flushing now", currentDocCount);
} else {
logger.info("--> [iteration {}] refreshing index", currentDocCount);
client().admin().indices().prepareRefresh(indexName).get();
}
}
logger.info("Completed ingestion of {} docs", currentDocCount);
}
});
}
Expand Down Expand Up @@ -218,4 +227,21 @@ public void stopShardRebalancing() {
.get()
);
}

public ClusterHealthStatus waitForRelocation() {
ClusterHealthRequest request = Requests.clusterHealthRequest()
.waitForNoRelocatingShards(true)
.timeout(TimeValue.timeValueSeconds(60))
.waitForEvents(Priority.LANGUID);
ClusterHealthResponse actionGet = client().admin().cluster().health(request).actionGet();
if (actionGet.isTimedOut()) {
logger.info(
"waitForRelocation timed out, cluster state:\n{}\n{}",
client().admin().cluster().prepareState().get().getState(),
client().admin().cluster().preparePendingClusterTasks().get()
);
assertThat("timed out waiting for relocation", actionGet.isTimedOut(), equalTo(false));
}
return actionGet.getStatus();
}
}
23 changes: 2 additions & 21 deletions ...rc/internalClusterTest/java/org/opensearch/remotemigration/RemotePrimaryRelocationIT.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -99,16 +99,7 @@ public void testRemotePrimaryRelocation() throws Exception {
.add(new MoveAllocationCommand("test", 0, primaryNodeName("test"), remoteNode))
.execute()
.actionGet();
ClusterHealthResponse clusterHealthResponse = client().admin()
.cluster()
.prepareHealth()
.setTimeout(TimeValue.timeValueSeconds(60))
.setWaitForEvents(Priority.LANGUID)
.setWaitForNoRelocatingShards(true)
.execute()
.actionGet();

assertEquals(0, clusterHealthResponse.getRelocatingShards());
waitForRelocation();
assertEquals(remoteNode, primaryNodeName("test"));
logger.info("--> relocation from docrep to remote complete");

Expand All @@ -123,16 +114,7 @@ public void testRemotePrimaryRelocation() throws Exception {
.add(new MoveAllocationCommand("test", 0, remoteNode, remoteNode2))
.execute()
.actionGet();
clusterHealthResponse = client().admin()
.cluster()
.prepareHealth()
.setTimeout(TimeValue.timeValueSeconds(60))
.setWaitForEvents(Priority.LANGUID)
.setWaitForNoRelocatingShards(true)
.execute()
.actionGet();

assertEquals(0, clusterHealthResponse.getRelocatingShards());
waitForRelocation();
assertEquals(remoteNode2, primaryNodeName("test"));

logger.info("--> relocation from remote to remote complete");
Expand All @@ -155,7 +137,6 @@ public void testRemotePrimaryRelocation() throws Exception {

public void testMixedModeRelocation_RemoteSeedingFail() throws Exception {
String docRepNode = internalCluster().startNode();
Client client = internalCluster().client(docRepNode);
ClusterUpdateSettingsRequest updateSettingsRequest = new ClusterUpdateSettingsRequest();
updateSettingsRequest.persistentSettings(Settings.builder().put(REMOTE_STORE_COMPATIBILITY_MODE_SETTING.getKey(), "mixed"));
assertAcked(client().admin().cluster().updateSettings(updateSettingsRequest).actionGet());
Expand Down
2 changes: 1 addition & 1 deletion server/src/main/java/org/opensearch/index/shard/IndexShard.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -2155,7 +2155,7 @@ public void waitForRemoteStoreSync(Runnable onProgress) throws IOException {
segmentUploadeCount = directory.getSegmentsUploadedToRemoteStore().size();
}
try {
Thread.sleep(TimeValue.timeValueSeconds(30).seconds());
Thread.sleep(TimeValue.timeValueSeconds(30).millis());
} catch (InterruptedException ie) {
throw new OpenSearchException("Interrupted waiting for completion of [{}]", ie);
}
Expand Down
Loading