-
Notifications
You must be signed in to change notification settings - Fork 1.9k
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.
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
Adding onNewCheckpoint to Start Replication on Replica Shard when Segment Replication is turned on #3540
Adding onNewCheckpoint to Start Replication on Replica Shard when Segment Replication is turned on #3540
Conversation
… latestcheckpoint and replaying logic is removed from this commit and will be added in a different PR Signed-off-by: Rishikesh1159 <[email protected]>
@@ -173,7 +173,7 @@ public final EngineConfig config() { | |||
* Return the latest active SegmentInfos from the engine. | |||
* @return {@link SegmentInfos} | |||
*/ | |||
protected abstract SegmentInfos getLatestSegmentInfos(); | |||
public abstract SegmentInfos getLatestSegmentInfos(); |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I don't think we need this additional method to fetch SegmentInfos, we already have getSegmentInfosSnapshot
on engine. Engine provides a default implementation, so we'll need to override this in NRTReplicationEngine.
Edited - I was wrong here, we still need this function but I don't think it needs to be public - public access to the infos can be through the snapshot method.
@@ -159,6 +160,7 @@ public class RecoverySettings { | |||
|
|||
private volatile ByteSizeValue chunkSize = DEFAULT_CHUNK_SIZE; | |||
|
|||
@Inject |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I don't think we need these Injects for SegmentReplicationTargetService, we can bind that class similar to PeerRecoverySourceService
in Node.java. We can create a SegmentReplicationSourceFactory there too and avoid injecting it.
b.bind(SegmentReplicationTargetService.class)
.toInstance(new SegmentReplicationTargetService(transportService, indicesService, recoverySettings, new SegmentReplicationSourceFactory(...)));
server/src/main/java/org/opensearch/indices/replication/SegmentReplicationTargetService.java
Outdated
Show resolved
Hide resolved
*/ | ||
private boolean shouldProcessCheckpoint(ReplicationCheckpoint requestCheckpoint, IndexShard indexshard) { | ||
if (indexshard.getState().equals(IndexShardState.STARTED) == false) { | ||
logger.debug("Ignore - shard is not started {}", indexshard.getState()); |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
nit - Can we be a bit more descriptive in these logs "Ignoring new replication checkpoint - ..."
I think these can also be trace logs?
return false; | ||
} | ||
ReplicationCheckpoint localCheckpoint = indexshard.getLatestReplicationCheckpoint(); | ||
logger.debug("Local Checkpoint {}", indexshard.getLatestReplicationCheckpoint()); |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
nit - I don't think we need to log this here, you are already logging the localCheckpoint below if validation fails.
*/ | ||
public ReplicationCheckpoint getLatestReplicationCheckpoint() { | ||
return new ReplicationCheckpoint(shardId, 0, 0, 0, 0); | ||
public IndexShardState getState() { |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I think this method is unnecessary we already have indexShard.state()
?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Got it. Will remove this in next commit
Signed-off-by: Rishikesh1159 <[email protected]>
Signed-off-by: Rishikesh1159 <[email protected]>
logger.trace("Checkpoint received {}", () -> requestCheckpoint); | ||
if (shouldProcessCheckpoint(requestCheckpoint, indexShard)) { | ||
logger.trace("Processing new checkpoint {}", requestCheckpoint); | ||
startReplication(requestCheckpoint, indexShard, new SegmentReplicationListener() { | ||
@Override | ||
public void onReplicationDone(SegmentReplicationState state) { | ||
logger.trace("Replication complete to {}", indexShard.getLatestReplicationCheckpoint()); | ||
} | ||
|
||
@Override | ||
public void onReplicationFailure(SegmentReplicationState state, OpenSearchException e, boolean sendShardFailure) { | ||
if (sendShardFailure == true) { | ||
indexShard.failShard("replication failure", e); | ||
} | ||
} | ||
}); | ||
|
||
} |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Lets remove trace logging if thats not needed add error logging on shard failure?
* @return true if checkpoint should be processed | ||
*/ | ||
private boolean shouldProcessCheckpoint(ReplicationCheckpoint requestCheckpoint, IndexShard indexShard) { | ||
if (indexShard.state().equals(IndexShardState.STARTED) == false) { |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
For my understanding what happens to a shard in RECOVERING state?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
We will take care of RECOVERING state logic in our next PR. Shards in RECOVERING state will never receive these new checkpoints. We'll be updating the recovery process to first ping the primary to have the local shard tracked, and then force a replication event
private boolean shouldProcessCheckpoint(ReplicationCheckpoint requestCheckpoint, IndexShard indexShard) { | ||
if (indexShard.state().equals(IndexShardState.STARTED) == false) { | ||
logger.trace("Ignoring new replication checkpoint - shard is not started {}", indexShard.state()); | ||
return false; | ||
} | ||
ReplicationCheckpoint localCheckpoint = indexShard.getLatestReplicationCheckpoint(); | ||
if (onGoingReplications.isShardReplicating(indexShard.shardId())) { | ||
logger.trace("Ignoring new replication checkpoint - shard is currently replicating to a checkpoint"); | ||
return false; | ||
} | ||
if (localCheckpoint.isAheadOf(requestCheckpoint)) { | ||
logger.trace( | ||
"Ignoring new replication checkpoint - Shard is already on checkpoint {} that is ahead of {}", | ||
localCheckpoint, | ||
requestCheckpoint | ||
); | ||
return false; | ||
} | ||
if (localCheckpoint.equals(requestCheckpoint)) { | ||
logger.trace("Ignoring new replication checkpoint - Shard is already on checkpoint {}", requestCheckpoint); | ||
return false; | ||
} | ||
return true; | ||
} |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Should this method be in IndexShard
instead?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Yes, thanks for catching this. I have moved this to IndexShard in new commit
…ce logs Signed-off-by: Rishikesh1159 <[email protected]>
Signed-off-by: Rishikesh1159 <[email protected]>
start gradle check |
Signed-off-by: Rishikesh1159 <[email protected]>
…1159/OpenSearch into seg-rep/new-checkpoint
Signed-off-by: Rishikesh1159 <[email protected]>
/** | ||
* Returns the lastest Replication Checkpoint that shard received | ||
*/ | ||
public ReplicationCheckpoint getLatestReplicationCheckpoint() { | ||
return new ReplicationCheckpoint(shardId, 0, 0, 0, 0); | ||
final SegmentInfos latestSegmentInfos = getLatestSegmentInfos(); | ||
return new ReplicationCheckpoint( |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
For a NRTReplicationEngine that is starting up, it will not yet have anything returned from getLatestSegmentInfos
until the first replication event completes. I think we need to null check here and return an "empty" ReplicationCheckpoint in that case, where gen and version could be SequenceNumbers.NO_OPS_PERFORMED
.
* @param requestCheckpoint received checkpoint that is checked for processing | ||
* @return true if checkpoint should be processed | ||
*/ | ||
public boolean shouldProcessCheckpoint(ReplicationCheckpoint requestCheckpoint) { |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
nit - this method can be final.
* @param requestCheckpoint received checkpoint that is checked for processing | ||
* @param indexShard replica shard on which checkpoint is received | ||
*/ | ||
public synchronized void onNewCheckpoint(final ReplicationCheckpoint requestCheckpoint, final IndexShard indexShard) { |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
If this class is no longer final, this method can be.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
This is still not possible as shardOperationOnReplica() in PublishCheckpointAction still needs to assert onNewCheckpoint(). Possibly need to use mockito-inline for making this or entire class final
server/src/main/java/org/opensearch/indices/replication/SegmentReplicationTargetService.java
Show resolved
Hide resolved
@@ -932,6 +934,15 @@ protected Node( | |||
.toInstance(new PeerRecoverySourceService(transportService, indicesService, recoverySettings)); | |||
b.bind(PeerRecoveryTargetService.class) | |||
.toInstance(new PeerRecoveryTargetService(threadPool, transportService, recoverySettings, clusterService)); | |||
b.bind(SegmentReplicationTargetService.class) |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
if you are binding here, you do not need @Inject
on SegmentReplicationTargetService
's constructor.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Shouldn't this binding also be gated by the feature flag, like SegmentReplicationSourceService
below?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Yes removed @Inject
SegmentReplicationTargetService
's constructor. Yes this should also be gated by feature flag. But need some changes in binding SegmentReplicationCheckpointPublisher in IndicesModule, will add them in next commit.
server/src/main/java/org/opensearch/indices/replication/SegmentReplicationTargetService.java
Show resolved
Hide resolved
* Returns the lastest segmentInfos | ||
*/ | ||
public SegmentInfos getLatestSegmentInfos() { | ||
return getEngine().getSegmentInfosSnapshot().get(); |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I missed this before - getSegmentInfosSnapshot returns a GatedCloseable
, we should use try-with-resources.
try(final GatedCloseable<SegmentInfos> snapshot = getEngine().getSegmentInfosSnapshot()) {
return snapshot.get();
}
there is also a getSegmentInfosSnapshot
method in IndexShard, we don't have to fetch off the engine.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
+1 to the note above. How is this different from the existing getSegmentInfosSnapshot ?
OpenSearch/server/src/main/java/org/opensearch/index/shard/IndexShard.java
Lines 4029 to 4031 in 04e43a7
public GatedCloseable<SegmentInfos> getSegmentInfosSnapshot() { | |
return getEngine().getSegmentInfosSnapshot(); | |
} |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
It is same as getSegmentInfosSnapshot(). So we can call that directly
start gradle check |
* Returns the lastest segmentInfos | ||
*/ | ||
public SegmentInfos getLatestSegmentInfos() { | ||
return getEngine().getSegmentInfosSnapshot().get(); |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
+1 to the note above. How is this different from the existing getSegmentInfosSnapshot ?
OpenSearch/server/src/main/java/org/opensearch/index/shard/IndexShard.java
Lines 4029 to 4031 in 04e43a7
public GatedCloseable<SegmentInfos> getSegmentInfosSnapshot() { | |
return getEngine().getSegmentInfosSnapshot(); | |
} |
server/src/main/java/org/opensearch/indices/replication/SegmentReplicationTargetService.java
Outdated
Show resolved
Hide resolved
* @return true if shard is currently replicating | ||
*/ | ||
public boolean isShardReplicating(ShardId shardId) { | ||
return onGoingTargetEvents.values().stream().anyMatch(t -> t.indexShard.shardId().equals(shardId)); |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Nit-pick - this is a problem to be solved down the line:
isShardReplicating
seems like a frequent check, and constantly flattening the map to a stream of values may end up being performance-intensive. We should consider if we could make the key of the map the shardId
to speed up this check.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Yes we can do that. But we have to change it in multiple places, so thinking of doing that in different PR
@@ -932,6 +934,15 @@ protected Node( | |||
.toInstance(new PeerRecoverySourceService(transportService, indicesService, recoverySettings)); | |||
b.bind(PeerRecoveryTargetService.class) | |||
.toInstance(new PeerRecoveryTargetService(threadPool, transportService, recoverySettings, clusterService)); | |||
b.bind(SegmentReplicationTargetService.class) |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Shouldn't this binding also be gated by the feature flag, like SegmentReplicationSourceService
below?
...r/src/test/java/org/opensearch/indices/replication/SegmentReplicationTargetServiceTests.java
Outdated
Show resolved
Hide resolved
...rc/test/java/org/opensearch/indices/replication/checkpoint/PublishCheckpointActionTests.java
Outdated
Show resolved
Hide resolved
Signed-off-by: Rishikesh1159 <[email protected]>
Signed-off-by: Rishikesh1159 <[email protected]>
…ceTests Signed-off-by: Rishikesh1159 <[email protected]>
start gradle check |
Signed-off-by: Rishikesh1159 <[email protected]>
* | ||
* @opensearch.internal | ||
*/ | ||
public final static class CheckpointPublisherProvider implements Provider<SegmentReplicationCheckpointPublisher> { |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Nitpick - if we choose to keep this provider (see my comment above), then this class can be made private
and static
doesn't make sense here
@@ -281,7 +284,30 @@ protected void configure() { | |||
bind(RetentionLeaseBackgroundSyncAction.class).asEagerSingleton(); | |||
bind(RetentionLeaseSyncer.class).asEagerSingleton(); | |||
if (FeatureFlags.isEnabled(FeatureFlags.REPLICATION_TYPE)) { | |||
bind(SegmentReplicationCheckpointPublisher.class).asEagerSingleton(); | |||
bind(SegmentReplicationCheckpointPublisher.class).toProvider(CheckpointPublisherProvider.class).asEagerSingleton(); |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Since you have an else
bind statement, you don't need really need a Provider here, and you can retain the @Inject
constructor
Signed-off-by: Rishikesh1159 <[email protected]>
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
LGTM, thanks @Rishikesh1159
* Bump reactor-netty-core from 1.0.16 to 1.0.19 in /plugins/repository-azure (#3360) * Bump reactor-netty-core in /plugins/repository-azure Bumps [reactor-netty-core](https://github.com/reactor/reactor-netty) from 1.0.16 to 1.0.19. - [Release notes](https://github.com/reactor/reactor-netty/releases) - [Commits](reactor/reactor-netty@v1.0.16...v1.0.19) --- updated-dependencies: - dependency-name: io.projectreactor.netty:reactor-netty-core dependency-type: direct:production update-type: version-update:semver-patch ... Signed-off-by: dependabot[bot] <[email protected]> * Updating SHAs Signed-off-by: dependabot[bot] <[email protected]> Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com> Co-authored-by: dependabot[bot] <dependabot[bot]@users.noreply.github.com> * [Type removal] _type removal from mocked responses of scroll hit tests (#3377) Signed-off-by: Suraj Singh <[email protected]> * [Type removal] Remove _type deprecation from script and conditional processor (#3239) * [Type removal] Remove _type deprecation from script and conditional processor Signed-off-by: Suraj Singh <[email protected]> * Spotless check apply Signed-off-by: Suraj Singh <[email protected]> * [Type removal] Remove _type from _bulk yaml test, scripts, unused constants (#3372) * [Type removal] Remove redundant _type deprecation checks in bulk request Signed-off-by: Suraj Singh <[email protected]> * [Type removal] bulk yaml tests validating deprecation on _type and removal from scripts Signed-off-by: Suraj Singh <[email protected]> * Fix Lucene-snapshots repo for jdk 17. (#3396) Signed-off-by: Marc Handalian <[email protected]> * Replace internal usages of 'master' term in 'server/src/internalClusterTest' directory (#2521) Signed-off-by: Tianli Feng <[email protected]> * [REMOVE] Cleanup deprecated thread pool types (FIXED_AUTO_QUEUE_SIZE) (#3369) Signed-off-by: Andriy Redko <[email protected]> * [Type removal] _type removal from tests of yaml tests (#3406) * [Type removal] _type removal from tests of yaml tests Signed-off-by: Suraj Singh <[email protected]> * Fix spotless failures Signed-off-by: Suraj Singh <[email protected]> * Fix assertion failures Signed-off-by: Suraj Singh <[email protected]> * Fix assertion failures in DoSectionTests Signed-off-by: Suraj Singh <[email protected]> * Add release notes for version 2.0.0 (#3410) Signed-off-by: Rabi Panda <[email protected]> * [Upgrade] Lucene-9.2.0-snapshot-ba8c3a8 (#3416) Upgrades to latest snapshot of lucene 9.2.0 in preparation for GA release. Signed-off-by: Nicholas Walter Knize <[email protected]> * Fix release notes for 2.0.0-rc1 version (#3418) This change removes some old commits from the 2.0.0-rc1 release notes. These commits were already released as part of 1.x releases. Add back some missing type removal commits to the 2.0.0 release notes Signed-off-by: Rabi Panda <[email protected]> * Bump version 2.1 to Lucene 9.2 after upgrade (#3424) Bumps Version.V_2_1_0 lucene version to 9.2 after backporting upgrage. Signed-off-by: Nicholas Walter Knize <[email protected]> * Bump com.gradle.enterprise from 3.10 to 3.10.1 (#3425) Bumps com.gradle.enterprise from 3.10 to 3.10.1. --- updated-dependencies: - dependency-name: com.gradle.enterprise dependency-type: direct:production update-type: version-update:semver-patch ... Signed-off-by: dependabot[bot] <[email protected]> Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com> * Bump reactor-core from 3.4.17 to 3.4.18 in /plugins/repository-azure (#3427) Bumps [reactor-core](https://github.com/reactor/reactor-core) from 3.4.17 to 3.4.18. - [Release notes](https://github.com/reactor/reactor-core/releases) - [Commits](reactor/reactor-core@v3.4.17...v3.4.18) --- updated-dependencies: - dependency-name: io.projectreactor:reactor-core dependency-type: direct:production update-type: version-update:semver-patch ... Signed-off-by: dependabot[bot] <[email protected]> Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com> Co-authored-by: dependabot[bot] <dependabot[bot]@users.noreply.github.com> * Bump gax-httpjson from 0.101.0 to 0.103.1 in /plugins/repository-gcs (#3426) Bumps [gax-httpjson](https://github.com/googleapis/gax-java) from 0.101.0 to 0.103.1. - [Release notes](https://github.com/googleapis/gax-java/releases) - [Changelog](https://github.com/googleapis/gax-java/blob/main/CHANGELOG.md) - [Commits](https://github.com/googleapis/gax-java/commits) --- updated-dependencies: - dependency-name: com.google.api:gax-httpjson dependency-type: direct:production update-type: version-update:semver-minor ... Signed-off-by: dependabot[bot] <[email protected]> Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com> Co-authored-by: dependabot[bot] <dependabot[bot]@users.noreply.github.com> * [segment replication]Introducing common Replication interfaces for segment replication and recovery code paths (#3234) * RecoveryState inherits from ReplicationState + RecoveryTarget inherits from ReplicationTarget Signed-off-by: Poojita Raj <[email protected]> * Refactoring: mixedClusterVersion error fix + move Stage to ReplicationState Signed-off-by: Poojita Raj <[email protected]> * pull ReplicationListener into a top level class + add javadocs + address review comments Signed-off-by: Poojita Raj <[email protected]> * fix javadoc Signed-off-by: Poojita Raj <[email protected]> * review changes Signed-off-by: Poojita Raj <[email protected]> * Refactoring the hierarchy relationship between repl and recovery Signed-off-by: Poojita Raj <[email protected]> * style fix Signed-off-by: Poojita Raj <[email protected]> * move package common under replication Signed-off-by: Poojita Raj <[email protected]> * rename to replication Signed-off-by: Poojita Raj <[email protected]> * rename and doc changes Signed-off-by: Poojita Raj <[email protected]> * [Type removal] Remove type from BulkRequestParser (#3423) * [Type removal] Remove type handling in bulk request parser Signed-off-by: Suraj Singh <[email protected]> * [Type removal] Remove testTypesStillParsedForBulkMonitoring as it is no longer present in codebase Signed-off-by: Suraj Singh <[email protected]> * Adding CheckpointRefreshListener to trigger when Segment replication is turned on and Primary shard refreshes (#3108) * Intial PR adding classes and tests related to checkpoint publishing Signed-off-by: Rishikesh1159 <[email protected]> * Putting a Draft PR with all changes in classes. Testing is still not included in this commit. Signed-off-by: Rishikesh1159 <[email protected]> * Wiring up index shard to new engine, spotless apply and removing unnecessary tests and logs Signed-off-by: Rishikesh1159 <[email protected]> * Adding Unit test for checkpointRefreshListener Signed-off-by: Rishikesh1159 <[email protected]> * Applying spotless check Signed-off-by: Rishikesh1159 <[email protected]> * Fixing import statements * Signed-off-by: Rishikesh1159 <[email protected]> * removing unused constructor in index shard Signed-off-by: Rishikesh1159 <[email protected]> * Addressing comments from last commit Signed-off-by: Rishikesh1159 <[email protected]> * Adding package-info.java files for two new packages Signed-off-by: Rishikesh1159 <[email protected]> * Adding test for null checkpoint publisher and addreesing PR comments Signed-off-by: Rishikesh1159 <[email protected]> * Add docs for indexshardtests and remove shard.refresh Signed-off-by: Rishikesh1159 <[email protected]> * Add a new Engine implementation for replicas with segment replication enabled. (#3240) * Change fastForwardProcessedSeqNo method in LocalCheckpointTracker to persisted checkpoint. This change inverts fastForwardProcessedSeqNo to fastForwardPersistedSeqNo for use in Segment Replication. This is so that a Segrep Engine can match the logic of InternalEngine where the seqNo is incremented with each operation, but only persisted in the tracker on a flush. With Segment Replication we bump the processed number with each operation received index/delete/noOp, and invoke this method when we receive a new set of segments to bump the persisted seqNo. Signed-off-by: Marc Handalian <[email protected]> * Extract Translog specific engine methods into an abstract class. This change extracts translog specific methods to an abstract engine class so that other engine implementations can reuse translog logic. Signed-off-by: Marc Handalian <[email protected]> * Add a separate Engine implementation for replicas with segment replication enabled. This change adds a new engine intended to be used on replicas with segment replication enabled. This engine does not wire up an IndexWriter, but still writes all operations to a translog. The engine uses a new ReaderManager that refreshes from an externally provided SegmentInfos. Signed-off-by: Marc Handalian <[email protected]> * Fix spotless checks. Signed-off-by: Marc Handalian <[email protected]> * Fix :server:compileInternalClusterTestJava compilation. Signed-off-by: Marc Handalian <[email protected]> * Fix failing test naming convention check. Signed-off-by: Marc Handalian <[email protected]> * PR feedback. - Removed isReadOnlyReplica from overloaded constructor and added feature flag checks. - Updated log msg in NRTReplicationReaderManager - cleaned up store ref counting in NRTReplicationEngine. Signed-off-by: Marc Handalian <[email protected]> * Fix spotless check. Signed-off-by: Marc Handalian <[email protected]> * Remove TranslogAwareEngine and build translog in NRTReplicationEngine. Signed-off-by: Marc Handalian <[email protected]> * Fix formatting Signed-off-by: Marc Handalian <[email protected]> * Add missing translog methods to NRTEngine. Signed-off-by: Marc Handalian <[email protected]> * Remove persistent seqNo check from fastForwardProcessedSeqNo. Signed-off-by: Marc Handalian <[email protected]> * PR feedback. Signed-off-by: Marc Handalian <[email protected]> * Add test specific to translog trimming. Signed-off-by: Marc Handalian <[email protected]> * Javadoc check. Signed-off-by: Marc Handalian <[email protected]> * Add failEngine calls to translog methods in NRTReplicationEngine. Roll xlog generation on replica when a new commit point is received. Signed-off-by: Marc Handalian <[email protected]> * Rename master to cluster_manager in the XContent Parser of ClusterHealthResponse (#3432) Signed-off-by: Tianli Feng <[email protected]> * Bump hadoop-minicluster in /test/fixtures/hdfs-fixture (#3359) Bumps hadoop-minicluster from 3.3.2 to 3.3.3. --- updated-dependencies: - dependency-name: org.apache.hadoop:hadoop-minicluster dependency-type: direct:production update-type: version-update:semver-patch ... Signed-off-by: dependabot[bot] <[email protected]> Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com> * Bump avro from 1.10.2 to 1.11.0 in /plugins/repository-hdfs (#3358) * Bump avro from 1.10.2 to 1.11.0 in /plugins/repository-hdfs Bumps avro from 1.10.2 to 1.11.0. --- updated-dependencies: - dependency-name: org.apache.avro:avro dependency-type: direct:production update-type: version-update:semver-minor ... Signed-off-by: dependabot[bot] <[email protected]> * Updating SHAs Signed-off-by: dependabot[bot] <[email protected]> Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com> Co-authored-by: dependabot[bot] <dependabot[bot]@users.noreply.github.com> * Fix testSetAdditionalRolesCanAddDeprecatedMasterRole() by removing the initial assertion (#3441) Signed-off-by: Tianli Feng <[email protected]> * Replace internal usages of 'master' term in 'server/src/test' directory (#2520) * Replace the non-inclusive terminology "master" with "cluster manager" in code comments, internal variable/method/class names, in `server/src/test` directory. * Backwards compatibility is not impacted. * Add a new unit test `testDeprecatedMasterNodeFilter()` to validate using `master:true` or `master:false` can filter the node in [Cluster Stats](https://opensearch.org/docs/latest/opensearch/rest-api/cluster-stats/) API, after the `master` role is deprecated in PR #2424 Signed-off-by: Tianli Feng <[email protected]> * Removing unused method from TransportSearchAction (#3437) * Removing unused method from TransportSearchAction Signed-off-by: Ankit Jain <[email protected]> * Set term vector flags to false for ._index_prefix field (#1901). (#3119) * Set term vector flags to false for ._index_prefix field (#1901). Signed-off-by: Vesa Pehkonen <[email protected]> * Replaced the FieldType copy ctor with ctor for the prefix field and replaced setting the field type parameters with setIndexOptions(). (#1901) Signed-off-by: Vesa Pehkonen <[email protected]> * Added tests for term vectors. (#1901) Signed-off-by: Vesa Pehkonen <[email protected]> * Fixed code formatting error. Signed-off-by: Vesa Pehkonen <[email protected]> Co-authored-by: sdp <[email protected]> * [BUG] Fixing org.opensearch.monitor.os.OsProbeTests > testLogWarnCpuMessageOnlyOnes when cgroups are available but cgroup stats is not (#3448) Signed-off-by: Andriy Redko <[email protected]> * [Segment Replication] Add SegmentReplicationTargetService to orchestrate replication events. (#3439) * Add SegmentReplicationTargetService to orchestrate replication events. This change introduces boilerplate classes for Segment Replication and a target service to orchestrate replication events. It also includes two refactors of peer recovery components for reuse. 1. Rename RecoveryFileChunkRequest to FileChunkRequest and extract code to handle throttling into ReplicationTarget. 2. Extracts a component to execute retryable requests over the transport layer. Signed-off-by: Marc Handalian <[email protected]> * Code cleanup. Signed-off-by: Marc Handalian <[email protected]> * Make SegmentReplicationTargetService component final so that it can not be extended by plugins. Signed-off-by: Marc Handalian <[email protected]> * Bump azure-core-http-netty from 1.11.9 to 1.12.0 in /plugins/repository-azure (#3474) Bumps [azure-core-http-netty](https://github.com/Azure/azure-sdk-for-java) from 1.11.9 to 1.12.0. - [Release notes](https://github.com/Azure/azure-sdk-for-java/releases) - [Commits](Azure/azure-sdk-for-java@azure-core-http-netty_1.11.9...azure-core_1.12.0) --- updated-dependencies: - dependency-name: com.azure:azure-core-http-netty dependency-type: direct:production update-type: version-update:semver-minor ... Signed-off-by: dependabot[bot] <[email protected]> * Update to Apache Lucene 9.2 (#3477) Signed-off-by: Andriy Redko <[email protected]> * Bump protobuf-java from 3.20.1 to 3.21.1 in /plugins/repository-hdfs (#3472) Signed-off-by: dependabot[bot] <[email protected]> * [Upgrade] Lucene-9.3.0-snapshot-823df23 (#3478) Upgrades to latest snapshot of lucene 9.3.0. Signed-off-by: Nicholas Walter Knize <[email protected]> * Filter out invalid URI and HTTP method in the error message of no handler found for a REST request (#3459) Filter out invalid URI and HTTP method of a error message, which shown when there is no handler found for a REST request sent by user, so that HTML special characters <>&"' will not shown in the error message. The error message is return as mine-type `application/json`, which can't contain active (script) content, so it's not a vulnerability. Besides, no browsers are going to render as html when the mine-type is that. While the common security scanners will raise a false-positive alarm for having HTML tags in the response without escaping the HTML special characters, so the solution only aims to satisfy the code security scanners. Signed-off-by: Tianli Feng <[email protected]> * Support use of IRSA for repository-s3 plugin credentials (#3475) * Support use of IRSA for repository-s3 plugin credentials Signed-off-by: Andriy Redko <[email protected]> * Address code review comments Signed-off-by: Andriy Redko <[email protected]> * Address code review comments Signed-off-by: Andriy Redko <[email protected]> * Bump google-auth-library-oauth2-http from 0.20.0 to 1.7.0 in /plugins/repository-gcs (#3473) * Bump google-auth-library-oauth2-http in /plugins/repository-gcs Bumps google-auth-library-oauth2-http from 0.20.0 to 1.7.0. --- updated-dependencies: - dependency-name: com.google.auth:google-auth-library-oauth2-http dependency-type: direct:production update-type: version-update:semver-major ... Signed-off-by: dependabot[bot] <[email protected]> * Updating SHAs Signed-off-by: dependabot[bot] <[email protected]> * Use variable to define the version of dependency google-auth-library-java Signed-off-by: Tianli Feng <[email protected]> Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com> Co-authored-by: dependabot[bot] <dependabot[bot]@users.noreply.github.com> Co-authored-by: Tianli Feng <[email protected]> * [Segment Replication] Added source-side classes for orchestrating replication events (#3470) This change expands on the existing SegmentReplicationSource interface and its corresponding Factory class by introducing an implementation where the replication source is a primary shard (PrimaryShardReplicationSource). These code paths execute on the target. The primary shard implementation creates the requests to be send to the source/primary shard. Correspondingly, this change also defines two request classes for the GET_CHECKPOINT_INFO and GET_SEGMENT_FILES requests as well as an abstract superclass. A CopyState class has been introduced that captures point-in-time, file-level details from an IndexShard. This implementation mirrors Lucene's NRT CopyState implementation. Finally, a service class has been introduce for segment replication that runs on the source side (SegmentReplicationSourceService) which handles these two types of incoming requests. This includes private handler classes that house the logic to respond to these requests, with some functionality stubbed for now. The service class also uses a simple map to cache CopyState objects that would be needed by replication targets. Unit tests have been added/updated for all new functionality. Signed-off-by: Kartik Ganesh <[email protected]> * [Dependency upgrade] google-oauth-client to 1.33.3 (#3500) Signed-off-by: Suraj Singh <[email protected]> * move bash flag to set statement (#3494) Passing bash with flags to the first argument of /usr/bin/env requires its own flag to interpret it correctly. Rather than use `env -S` to split the argument, have the script `set -e` to enable the same behavior explicitly in preinst and postinst scripts. Also set `-o pipefail` for consistency. Closes: #3492 Signed-off-by: Cole White <[email protected]> * Support use of IRSA for repository-s3 plugin credentials: added YAML Rest test case (#3499) Signed-off-by: Andriy Redko <[email protected]> * Bump azure-storage-common from 12.15.0 to 12.16.0 in /plugins/repository-azure (#3517) * Bump azure-storage-common in /plugins/repository-azure Bumps [azure-storage-common](https://github.com/Azure/azure-sdk-for-java) from 12.15.0 to 12.16.0. - [Release notes](https://github.com/Azure/azure-sdk-for-java/releases) - [Commits](Azure/azure-sdk-for-java@azure-storage-blob_12.15.0...azure-storage-blob_12.16.0) --- updated-dependencies: - dependency-name: com.azure:azure-storage-common dependency-type: direct:production update-type: version-update:semver-minor ... Signed-off-by: dependabot[bot] <[email protected]> * Updating SHAs Signed-off-by: dependabot[bot] <[email protected]> Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com> Co-authored-by: dependabot[bot] <dependabot[bot]@users.noreply.github.com> * Bump google-oauth-client from 1.33.3 to 1.34.0 in /plugins/discovery-gce (#3516) * Bump google-oauth-client from 1.33.3 to 1.34.0 in /plugins/discovery-gce Bumps [google-oauth-client](https://github.com/googleapis/google-oauth-java-client) from 1.33.3 to 1.34.0. - [Release notes](https://github.com/googleapis/google-oauth-java-client/releases) - [Changelog](https://github.com/googleapis/google-oauth-java-client/blob/main/CHANGELOG.md) - [Commits](googleapis/google-oauth-java-client@v1.33.3...v1.34.0) --- updated-dependencies: - dependency-name: com.google.oauth-client:google-oauth-client dependency-type: direct:production update-type: version-update:semver-minor ... Signed-off-by: dependabot[bot] <[email protected]> * Updating SHAs Signed-off-by: dependabot[bot] <[email protected]> Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com> Co-authored-by: dependabot[bot] <dependabot[bot]@users.noreply.github.com> * Fix the support of RestClient Node Sniffer for version 2.x and update tests (#3487) Fix the support of RestClient Node Sniffer for OpenSearch 2.x, and update unit tests for OpenSearch. The current code contains the logic to be compatible with Elasticsearch 2.x version, which is conflict with OpenSearch 2.x, so removed that part of legacy code. * Update the script create_test_nodes_info.bash to dump the response of Nodes Info API GET _nodes/http for OpenSearch 1.0 and 2.0 version, which used for unit test. * Remove the support of Elasticsearch version 2.x for the Sniffer * Update unit test to validate the Sniffer compatible with OpenSearch 1.x and 2.x * Update the API response parser to meet the array notation (in ES 6.1 and above) for the node attributes setting. It will result the value of `node.attr` setting will not be parsed as array in the Sniffer, when using the Sniffer on cluster in Elasticsearch 6.0 and above. * Replace "master" node role with "cluster_manager" in unit test Signed-off-by: Tianli Feng <[email protected]> * Bump com.diffplug.spotless from 6.6.1 to 6.7.0 (#3513) Bumps com.diffplug.spotless from 6.6.1 to 6.7.0. --- updated-dependencies: - dependency-name: com.diffplug.spotless dependency-type: direct:production update-type: version-update:semver-minor ... Signed-off-by: dependabot[bot] <[email protected]> Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com> * Bump guava from 18.0 to 23.0 in /plugins/ingest-attachment (#3357) * Bump guava from 18.0 to 23.0 in /plugins/ingest-attachment Bumps [guava](https://github.com/google/guava) from 18.0 to 23.0. - [Release notes](https://github.com/google/guava/releases) - [Commits](google/guava@v18.0...v23.0) --- updated-dependencies: - dependency-name: com.google.guava:guava dependency-type: direct:production update-type: version-update:semver-major ... Signed-off-by: dependabot[bot] <[email protected]> * Updating SHAs Signed-off-by: dependabot[bot] <[email protected]> * Add more ingorance of using internal java API sun.misc.Unsafe Signed-off-by: Tianli Feng <[email protected]> Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com> Co-authored-by: dependabot[bot] <dependabot[bot]@users.noreply.github.com> Co-authored-by: Tianli Feng <[email protected]> * Added bwc version 2.0.1 (#3452) Signed-off-by: Kunal Kotwani <[email protected]> Co-authored-by: opensearch-ci-bot <[email protected]> * Add release notes for 1.3.3 (#3549) Signed-off-by: Xue Zhou <[email protected]> * [Upgrade] Lucene-9.3.0-snapshot-b7231bb (#3537) Upgrades to latest snapshot of lucene 9.3; including reducing maxFullFlushMergeWaitMillis in LuceneTest.testWrapLiveDocsNotExposeAbortedDocuments to 0 ms to ensure aborted docs are not merged away in the test with the new mergeOnRefresh default policy. Signed-off-by: Nicholas Walter Knize <[email protected]> * [Remote Store] Upload segments to remote store post refresh (#3460) * Add RemoteDirectory interface to copy segment files to/from remote store Signed-off-by: Sachin Kale <[email protected]> Co-authored-by: Sachin Kale <[email protected]> * Add index level setting for remote store Signed-off-by: Sachin Kale <[email protected]> Co-authored-by: Sachin Kale <[email protected]> * Add RemoteDirectoryFactory and use RemoteDirectory instance in RefreshListener Co-authored-by: Sachin Kale <[email protected]> Signed-off-by: Sachin Kale <[email protected]> * Upload segment to remote store post refresh Signed-off-by: Sachin Kale <[email protected]> Co-authored-by: Sachin Kale <[email protected]> * Fixing VerifyVersionConstantsIT test failure (#3574) Signed-off-by: Andriy Redko <[email protected]> * Bump jettison from 1.4.1 to 1.5.0 in /plugins/discovery-azure-classic (#3571) * Bump jettison from 1.4.1 to 1.5.0 in /plugins/discovery-azure-classic Bumps [jettison](https://github.com/jettison-json/jettison) from 1.4.1 to 1.5.0. - [Release notes](https://github.com/jettison-json/jettison/releases) - [Commits](jettison-json/jettison@jettison-1.4.1...jettison-1.5.0) --- updated-dependencies: - dependency-name: org.codehaus.jettison:jettison dependency-type: direct:production update-type: version-update:semver-minor ... Signed-off-by: dependabot[bot] <[email protected]> * Updating SHAs Signed-off-by: dependabot[bot] <[email protected]> Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com> Co-authored-by: dependabot[bot] <dependabot[bot]@users.noreply.github.com> * Bump google-api-services-storage from v1-rev20200814-1.30.10 to v1-rev20220608-1.32.1 in /plugins/repository-gcs (#3573) * Bump google-api-services-storage in /plugins/repository-gcs Bumps google-api-services-storage from v1-rev20200814-1.30.10 to v1-rev20220608-1.32.1. --- updated-dependencies: - dependency-name: com.google.apis:google-api-services-storage dependency-type: direct:production ... Signed-off-by: dependabot[bot] <[email protected]> * Updating SHAs Signed-off-by: dependabot[bot] <[email protected]> * Upgrade Google HTTP Client to 1.42.0 Signed-off-by: Xue Zhou <[email protected]> Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com> Co-authored-by: dependabot[bot] <dependabot[bot]@users.noreply.github.com> Co-authored-by: Xue Zhou <[email protected]> * Add flat_skew setting to node overload decider (#3563) * Add flat_skew setting to node overload decider Signed-off-by: Rishab Nahata <[email protected]> * Bump xmlbeans from 5.0.3 to 5.1.0 in /plugins/ingest-attachment (#3572) * Bump xmlbeans from 5.0.3 to 5.1.0 in /plugins/ingest-attachment Bumps xmlbeans from 5.0.3 to 5.1.0. --- updated-dependencies: - dependency-name: org.apache.xmlbeans:xmlbeans dependency-type: direct:production update-type: version-update:semver-minor ... Signed-off-by: dependabot[bot] <[email protected]> * Updating SHAs Signed-off-by: dependabot[bot] <[email protected]> Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com> Co-authored-by: dependabot[bot] <dependabot[bot]@users.noreply.github.com> * Bump google-oauth-client from 1.34.0 to 1.34.1 in /plugins/discovery-gce (#3570) * Bump google-oauth-client from 1.34.0 to 1.34.1 in /plugins/discovery-gce Bumps [google-oauth-client](https://github.com/googleapis/google-oauth-java-client) from 1.34.0 to 1.34.1. - [Release notes](https://github.com/googleapis/google-oauth-java-client/releases) - [Changelog](https://github.com/googleapis/google-oauth-java-client/blob/main/CHANGELOG.md) - [Commits](googleapis/google-oauth-java-client@v1.34.0...v1.34.1) --- updated-dependencies: - dependency-name: com.google.oauth-client:google-oauth-client dependency-type: direct:production update-type: version-update:semver-patch ... Signed-off-by: dependabot[bot] <[email protected]> * Updating SHAs Signed-off-by: dependabot[bot] <[email protected]> Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com> Co-authored-by: dependabot[bot] <dependabot[bot]@users.noreply.github.com> * Fix for bug showing incorrect awareness attributes count in AwarenessAllocationDecider (#3428) * Fix for bug showing incorrect awareness attributes count in AwarenessAllocationDecider Signed-off-by: Anshu Agarwal <[email protected]> * Added bwc version 1.3.4 (#3552) Signed-off-by: GitHub <[email protected]> Co-authored-by: opensearch-ci-bot <[email protected]> * Support dynamic node role (#3436) * Support unknown node role Currently OpenSearch only supports several built-in nodes like data node role. If specify unknown node role, OpenSearch node will fail to start. This limit how to extend OpenSearch to support some extension function. For example, user may prefer to run ML tasks on some dedicated node which doesn't serve as any built-in node roles. So the ML tasks won't impact OpenSearch core function. This PR removed the limitation and user can specify any node role and OpenSearch will start node correctly with that unknown role. This opens the door for plugin developer to run specific tasks on dedicated nodes. Issue: #2877 Signed-off-by: Yaliang Wu <[email protected]> * fix cat nodes rest API spec Signed-off-by: Yaliang Wu <[email protected]> * fix mixed cluster IT failure Signed-off-by: Yaliang Wu <[email protected]> * add DynamicRole Signed-off-by: Yaliang Wu <[email protected]> * change generator method name Signed-off-by: Yaliang Wu <[email protected]> * fix failed docker test Signed-off-by: Yaliang Wu <[email protected]> * transform role name to lower case to avoid confusion Signed-off-by: Yaliang Wu <[email protected]> * transform the node role abbreviation to lower case Signed-off-by: Yaliang Wu <[email protected]> * fix checkstyle Signed-off-by: Yaliang Wu <[email protected]> * add test for case-insensitive role name change Signed-off-by: Yaliang Wu <[email protected]> * Rename package 'o.o.action.support.master' to 'o.o.action.support.clustermanager' (#3556) * Rename package org.opensearch.action.support.master to org.opensearch.action.support.clustermanager Signed-off-by: Tianli Feng <[email protected]> * Rename classes with master term in the package org.opensearch.action.support.master Signed-off-by: Tianli Feng <[email protected]> * Deprecate classes in org.opensearch.action.support.master Signed-off-by: Tianli Feng <[email protected]> * Remove pakcage o.o.action.support.master Signed-off-by: Tianli Feng <[email protected]> * Move package-info back Signed-off-by: Tianli Feng <[email protected]> * Move package-info to new folder Signed-off-by: Tianli Feng <[email protected]> * Correct the package-info Signed-off-by: Tianli Feng <[email protected]> * Fixing flakiness of ShuffleForcedMergePolicyTests (#3591) Signed-off-by: Andriy Redko <[email protected]> * Deprecate classes in org.opensearch.action.support.master (#3593) Signed-off-by: Tianli Feng <[email protected]> * Add release notes for version 2.0.1 (#3595) Signed-off-by: Kunal Kotwani <[email protected]> * Fix NPE when minBound/maxBound is not set before being called. (#3605) Signed-off-by: George Apaaboah <[email protected]> * Added bwc version 2.0.2 (#3613) Co-authored-by: opensearch-ci-bot <[email protected]> * Fix false positive query timeouts due to using cached time (#3454) * Fix false positive query timeouts due to using cached time Signed-off-by: Ahmad AbuKhalil <[email protected]> * delegate nanoTime call to SearchContext Signed-off-by: Ahmad AbuKhalil <[email protected]> * add override to SearchContext getRelativeTimeInMillis to force non cached time Signed-off-by: Ahmad AbuKhalil <[email protected]> * Fix random gradle check failure issue 3584. (#3627) * [Segment Replication] Add components for segment replication to perform file copy. (#3525) * Add components for segment replication to perform file copy. This change adds the required components to SegmentReplicationSourceService to initiate copy and react to lifecycle events. Along with new components it refactors common file copy code from RecoverySourceHandler into reusable pieces. Signed-off-by: Marc Handalian <[email protected]> * Deprecate public methods and variables with master term in package 'org.opensearch.action.support.master' (#3617) Signed-off-by: Tianli Feng <[email protected]> * Add replication orchestration for a single shard (#3533) * implement segment replication target Signed-off-by: Poojita Raj <[email protected]> * test added Signed-off-by: Poojita Raj <[email protected]> * changes to tests + finalizeReplication Signed-off-by: Poojita Raj <[email protected]> * fix style check Signed-off-by: Poojita Raj <[email protected]> * addressing comments + fix gradle check Signed-off-by: Poojita Raj <[email protected]> * added test + addressed review comments Signed-off-by: Poojita Raj <[email protected]> * [BUG] opensearch crashes on closed client connection before search reply (#3626) * [BUG] opensearch crashes on closed client connection before search reply Signed-off-by: Andriy Redko <[email protected]> * Addressing code review comments Signed-off-by: Andriy Redko <[email protected]> * Add all deprecated method in the package with new name 'org.opensearch.action.support.clustermanager' (#3644) Signed-off-by: Tianli Feng <[email protected]> * Introduce TranslogManager implementations decoupled from the Engine (#3638) * Introduce decoupled translog manager interfaces Signed-off-by: Bukhtawar Khan <[email protected]> * Adding onNewCheckpoint to Start Replication on Replica Shard when Segment Replication is turned on (#3540) * Adding onNewCheckpoint and it's test to start replication. SCheck for latestcheckpoint and replaying logic is removed from this commit and will be added in a different PR Signed-off-by: Rishikesh1159 <[email protected]> * Changing binding/inject logic and addressing comments from PR Signed-off-by: Rishikesh1159 <[email protected]> * Applying spotless check Signed-off-by: Rishikesh1159 <[email protected]> * Moving shouldProcessCheckpoint() to IndexShard, and removing some trace logs Signed-off-by: Rishikesh1159 <[email protected]> * applying spotlessApply Signed-off-by: Rishikesh1159 <[email protected]> * Adding more info to log statement in targetservice class Signed-off-by: Rishikesh1159 <[email protected]> * applying spotlessApply Signed-off-by: Rishikesh1159 <[email protected]> * Addressing comments on PR Signed-off-by: Rishikesh1159 <[email protected]> * Adding teardown() in SegmentReplicationTargetServiceTests. Signed-off-by: Rishikesh1159 <[email protected]> * fixing testShouldProcessCheckpoint() in SegmentReplicationTargetServiceTests Signed-off-by: Rishikesh1159 <[email protected]> * Removing CheckpointPublisherProvider in IndicesModule Signed-off-by: Rishikesh1159 <[email protected]> * spotless check apply Signed-off-by: Rishikesh1159 <[email protected]> * Remove class org.opensearch.action.support.master.AcknowledgedResponse (#3662) * Remove class org.opensearch.action.support.master.AcknowledgedResponse Signed-off-by: Tianli Feng <[email protected]> * Remove class org.opensearch.action.support.master.AcknowledgedRequest RequestBuilder ShardsAcknowledgedResponse Signed-off-by: Tianli Feng <[email protected]> * Restore AcknowledgedResponse and AcknowledgedRequest to package org.opensearch.action.support.master (#3669) Signed-off-by: Tianli Feng <[email protected]> * [BUG] Custom POM configuration for ZIP publication produces duplicit tags (url, scm) (#3656) * [BUG] Custom POM configuration for ZIP publication produces duplicit tags (url, scm) Signed-off-by: Andriy Redko <[email protected]> * Added test case for pluginZip with POM Signed-off-by: Andriy Redko <[email protected]> * Support both Gradle 6.8.x and Gradle 7.4.x Signed-off-by: Andriy Redko <[email protected]> * Adding 2.2.0 Bwc version to main (#3673) * Upgraded to t-digest 3.3. (#3634) * Revert renaming method onMaster() and offMaster() in interface LocalNodeMasterListener (#3686) Signed-off-by: Tianli Feng <[email protected]> * Upgrading AWS SDK dependency for native plugins (#3694) * Merge branch 'feature/point_in_time' of https://github.com/opensearch-project/OpenSearch into fb Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com> Co-authored-by: dependabot[bot] <dependabot[bot]@users.noreply.github.com> Co-authored-by: Suraj Singh <[email protected]> Co-authored-by: Marc Handalian <[email protected]> Co-authored-by: Tianli Feng <[email protected]> Co-authored-by: Andriy Redko <[email protected]> Co-authored-by: Rabi Panda <[email protected]> Co-authored-by: Nick Knize <[email protected]> Co-authored-by: Poojita Raj <[email protected]> Co-authored-by: Rishikesh Pasham <[email protected]> Co-authored-by: Ankit Jain <[email protected]> Co-authored-by: vpehkone <[email protected]> Co-authored-by: sdp <[email protected]> Co-authored-by: Kartik Ganesh <[email protected]> Co-authored-by: Cole White <[email protected]> Co-authored-by: opensearch-trigger-bot[bot] <98922864+opensearch-trigger-bot[bot]@users.noreply.github.com> Co-authored-by: opensearch-ci-bot <[email protected]> Co-authored-by: Xue Zhou <[email protected]> Co-authored-by: Sachin Kale <[email protected]> Co-authored-by: Sachin Kale <[email protected]> Co-authored-by: Xue Zhou <[email protected]> Co-authored-by: Rishab Nahata <[email protected]> Co-authored-by: Anshu Agarwal <[email protected]> Co-authored-by: Yaliang Wu <[email protected]> Co-authored-by: Kunal Kotwani <[email protected]> Co-authored-by: George Apaaboah <[email protected]> Co-authored-by: Ahmad AbuKhalil <[email protected]> Co-authored-by: Bukhtawar Khan <[email protected]> Co-authored-by: Sarat Vemulapalli <[email protected]> Co-authored-by: Daniel (dB.) Doubrovkine <[email protected]>
…ment Replication is turned on (#3540) * Adding onNewCheckpoint and it's test to start replication. SCheck for latestcheckpoint and replaying logic is removed from this commit and will be added in a different PR Signed-off-by: Rishikesh1159 <[email protected]> * Changing binding/inject logic and addressing comments from PR Signed-off-by: Rishikesh1159 <[email protected]> * Applying spotless check Signed-off-by: Rishikesh1159 <[email protected]> * Moving shouldProcessCheckpoint() to IndexShard, and removing some trace logs Signed-off-by: Rishikesh1159 <[email protected]> * applying spotlessApply Signed-off-by: Rishikesh1159 <[email protected]> * Adding more info to log statement in targetservice class Signed-off-by: Rishikesh1159 <[email protected]> * applying spotlessApply Signed-off-by: Rishikesh1159 <[email protected]> * Addressing comments on PR Signed-off-by: Rishikesh1159 <[email protected]> * Adding teardown() in SegmentReplicationTargetServiceTests. Signed-off-by: Rishikesh1159 <[email protected]> * fixing testShouldProcessCheckpoint() in SegmentReplicationTargetServiceTests Signed-off-by: Rishikesh1159 <[email protected]> * Removing CheckpointPublisherProvider in IndicesModule Signed-off-by: Rishikesh1159 <[email protected]> * spotless check apply Signed-off-by: Rishikesh1159 <[email protected]>
…ment Replication is turned on (opensearch-project#3540) * Adding onNewCheckpoint and it's test to start replication. SCheck for latestcheckpoint and replaying logic is removed from this commit and will be added in a different PR Signed-off-by: Rishikesh1159 <[email protected]> * Changing binding/inject logic and addressing comments from PR Signed-off-by: Rishikesh1159 <[email protected]> * Applying spotless check Signed-off-by: Rishikesh1159 <[email protected]> * Moving shouldProcessCheckpoint() to IndexShard, and removing some trace logs Signed-off-by: Rishikesh1159 <[email protected]> * applying spotlessApply Signed-off-by: Rishikesh1159 <[email protected]> * Adding more info to log statement in targetservice class Signed-off-by: Rishikesh1159 <[email protected]> * applying spotlessApply Signed-off-by: Rishikesh1159 <[email protected]> * Addressing comments on PR Signed-off-by: Rishikesh1159 <[email protected]> * Adding teardown() in SegmentReplicationTargetServiceTests. Signed-off-by: Rishikesh1159 <[email protected]> * fixing testShouldProcessCheckpoint() in SegmentReplicationTargetServiceTests Signed-off-by: Rishikesh1159 <[email protected]> * Removing CheckpointPublisherProvider in IndicesModule Signed-off-by: Rishikesh1159 <[email protected]> * spotless check apply Signed-off-by: Rishikesh1159 <[email protected]>
Signed-off-by: Rishikesh1159 <[email protected]>
Signed-off-by: Rishikesh1159 <[email protected]>
…#3943 #3963 From main branch (#4181) * Resolving import conflict in Node.java and mergining PR #3525. Signed-off-by: Rishikesh1159 <[email protected]> * Resolving conflicts and merging PR #3533. Signed-off-by: Rishikesh1159 <[email protected]> * Resolving conflicts and Merging PR #3540. Signed-off-by: Rishikesh1159 <[email protected]> * Applying spotlesscheck and fixing wildcard imports. Signed-off-by: Rishikesh1159 <[email protected]> * [Segment Replication] Fixing flaky test failure happening for testShardAlreadyReplicating() (#3943) * Fixing flaky test failure happening for testShardAlreadyReplicating() Signed-off-by: Rishikesh1159 <[email protected]> * Fix possible flaky test for testBeforeIndexShardClosed_CancelsOngoingReplications() (#3963) * Fixing flaky test failure happening for testShardAlreadyReplicating() Signed-off-by: Rishikesh1159 <[email protected]> * Removing assert segrep() in getProcessedLocalCheckpoint() of Index.shard class. Signed-off-by: Rishikesh1159 <[email protected]> * Adding back assert statement and make index setting to segment replication in SegmentReplicationSourceHandlerTests and SegmentReplicationTargetServiceTests. Signed-off-by: Rishikesh1159 <[email protected]> * Revert "Adding back assert statement and make index setting to segment replication in SegmentReplicationSourceHandlerTests and SegmentReplicationTargetServiceTests." Signed-off-by: Rishikesh1159 <[email protected]> This reverts commit 8c5753b. Signed-off-by: Rishikesh1159 <[email protected]> Co-authored-by: Marc Handalian <[email protected]> Co-authored-by: Poojita Raj <[email protected]>
Signed-off-by: Rishikesh1159 [email protected]
Description
This PR will allow us to start replication on replica shard when a new checkpoint is received. The logic of replaying new checkpoints and storing Latest received checkpoint which comes under same issue 3110, but will be implemented later in a different PR. Some changes like making SegmentReplicationTargetService class not final and Injecting SegmentReplicationTargetService needs discussion/opinion from other developers, please drop your thoughts on these changes also.
This is a part of the process of merging our feature branch - feature/segment-replication - back into main by re-PRing our changes from the feature branch.
The breakdown of the plan to merge to main is detailed here: #2355
For added context on segment replication - here's the design proposal #2229
Issues Resolved
#3110
Check List
By submitting this pull request, I confirm that my contribution is made under the terms of the Apache 2.0 license.
For more information on following Developer Certificate of Origin and signing off your commits, please check here.