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.

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

[Remote Routing Table]Read remote index routing #13894

Conversation

Arpit-Bandejiya
Copy link
Contributor

Description

This is an draft PR for remote index routing read. The PR is dependent on cluster state service publication PR #13835 and initial remote routing setup: #13304

Related Issues

Resolves #[Issue number to be closed when this PR is merged]

Check List

  • New functionality includes testing.
    • All tests pass
  • New functionality has been documented.
    • New functionality has javadoc added
  • API changes companion pull request created.
  • Failing checks are inspected and point to the corresponding known issue(s) (See: Troubleshooting Failing Builds)
  • Commits are signed per the DCO using --signoff
  • Commit changes are listed out in CHANGELOG.md file (See: Changelog)
  • Public documentation issue/PR created

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.

@Arpit-Bandejiya Arpit-Bandejiya changed the title Read remote index routing [Remote Routing Table]Read remote index routing May 30, 2024
}).collect(Collectors.toList());
}

public static List<String> getIndicesRoutingDeleted(RoutingTable previousRoutingTable, RoutingTable currentRoutingTable) {
Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This function is used for getting the deleted indices for creating diff in manifest

return deletedIndicesRouting;
}

public static List<String> getIndicesRoutingUpdated(RoutingTable previousRoutingTable, RoutingTable currentRoutingTable) {
Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This function is used for getting the updated index routing for creating diff in manifest.

return null;
}

public List<ClusterMetadataManifest.UploadedIndexMetadata> getUpdatedIndexRoutingTableMetadata(List<String> updatedIndicesRouting, List<ClusterMetadataManifest.UploadedIndexMetadata> allIndicesRouting) {
Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Used to create the Metadata from the list of updated indicesRouting from Manifest in read flow.

Copy link
Contributor

❌ Gradle check result for 933933c: FAILURE

Please examine the workflow log, locate, and copy-paste the failure(s) below, then iterate to green. Is the failure a flaky test unrelated to your change?

Comment on lines 145 to 156
} catch (IOException e) {
logger.info("RoutingTable read failed with error: {}", e.toString());
}
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

throw exception here

Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

use the correct exception logging

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thowed RemoteStateTransfterException.

} catch (IOException e) {
logger.info("RoutingTable read failed with error: {}", e.toString());
}
return null;
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Lets throw exception in case of failures, returning null can generate NPEs in the request flow.

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Done.

Comment on lines 170 to 184
public static List<String> getIndicesRoutingUpdated(RoutingTable previousRoutingTable, RoutingTable currentRoutingTable) {
List<String> updatedIndicesRouting = new ArrayList<>();
for(IndexRoutingTable currentIndicesRouting: currentRoutingTable.getIndicesRouting().values()) {
if(!previousRoutingTable.getIndicesRouting().containsKey(currentIndicesRouting.getIndex().getName())) {
// Latest Routing Table does not have entry for the index which means the index is created
updatedIndicesRouting.add(currentIndicesRouting.getIndex().getName());
} else {
if(previousRoutingTable.getIndicesRouting().get(currentIndicesRouting.getIndex().getName()).equals(currentIndicesRouting)) {
// if the latest routing table has the same routing table as the previous routing table, then the index is not updated
continue;
}
updatedIndicesRouting.add(currentIndicesRouting.getIndex().getName());
}
}
return updatedIndicesRouting;
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Can we reuse DiffableUtils rather than replicating the logic here

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Done.

Comment on lines +106 to +121
private String getIndexRoutingFileName() {
return String.join(
DELIMITER,
INDEX_ROUTING_FILE_PREFIX,
RemoteStoreUtils.invertLong(System.currentTimeMillis())
);

}
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

cc: @ashking94 for reviewing prefix changes

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This function is not in use for read flow, got added by mistake. I think the following PR has changes related to the prefix hashing: https://github.com/opensearch-project/OpenSearch/pull/13870/files#diff-54b9eb2094b2f28930e8c235f57f00d8cb6547e790684cf54e533756e1251f85R216

Comment on lines 123 to 137
return () -> readAsync(
blobContainer,
blobFileName,
threadPool.executor(ThreadPool.Names.GENERIC),
ActionListener.wrap(response -> latchedActionListener.onResponse(new RemoteIndexRoutingResult(index.getName(), response.readIndexRoutingTable(index))), latchedActionListener::onFailure)
);
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Lets use a dedicated thread pool for remote state interactions

Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Where can I find the reference for RemoteIndexRoutingResult ?

*
* @opensearch.internal
*/
public class RemoteRoutingTableService implements Closeable {
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Would prefer AbstractLifecycleComponent to be implemented

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This is taken care in #13577


}

public CheckedRunnable<IOException> getAsyncIndexMetadataReadAction(
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Where can I see the consumer of this function?

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

blobStoreRepository = (BlobStoreRepository) repository;
}

public static class RemoteIndexRoutingResult {
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Why is this class needed? IndexRoutingTable already holds Index info

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The idea was to keep something similar to RemoteResult type of object. For now, I have removed it. We can take an call when the parent PRs are merged.

public static List<String> getIndicesRoutingUpdated(RoutingTable previousRoutingTable, RoutingTable currentRoutingTable) {
List<String> updatedIndicesRouting = new ArrayList<>();
for(IndexRoutingTable currentIndicesRouting: currentRoutingTable.getIndicesRouting().values()) {
if(!previousRoutingTable.getIndicesRouting().containsKey(currentIndicesRouting.getIndex().getName())) {
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Don't we need to compare versions?

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Do you mean when the previsionRoutingTable of older version of manifest and we are on the latest one?

Copy link
Contributor

github-actions bot commented Jun 3, 2024

❌ Gradle check result for a7fb428: FAILURE

Please examine the workflow log, locate, and copy-paste the failure(s) below, then iterate to green. Is the failure a flaky test unrelated to your change?

Copy link
Contributor

github-actions bot commented Jun 4, 2024

❌ Gradle check result for 129b1d3: FAILURE

Please examine the workflow log, locate, and copy-paste the failure(s) below, then iterate to green. Is the failure a flaky test unrelated to your change?

Signed-off-by: Arpit Bandejiya <abandeji@amazon.com>
Copy link
Contributor

github-actions bot commented Jun 6, 2024

❌ Gradle check result for 768585e: FAILURE

Please examine the workflow log, locate, and copy-paste the failure(s) below, then iterate to green. Is the failure a flaky test unrelated to your change?

boolean readTemplatesMetadata,
boolean readDiscoveryNodes,
boolean readClusterBlocks,
List<UploadedIndexMetadata> indicesRoutingToRead,
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

How are we getting this list indicesRoutingToRead

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Added.

}
}

public List<ClusterMetadataManifest.UploadedIndexMetadata> getUpdatedIndexRoutingTableMetadata(List<String> updatedIndicesRouting, List<ClusterMetadataManifest.UploadedIndexMetadata> allIndicesRouting) {
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Where is this getting called from in RemoteClusterStateService?

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Added in the flow now.

Comment on lines +926 to +932
ClusterState.Builder clusterStateBuilder = ClusterState.builder(previousState);
Map<String, IndexRoutingTable> indicesRouting = new HashMap<>(previousState.routingTable().getIndicesRouting());


readIndexRoutingTableResults.forEach(indexRoutingTable -> {
indicesRouting.put(indexRoutingTable.getIndex().getName(), indexRoutingTable);
});
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

What if there are index in previousState which are now deleted?

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Covered in the main call, added the flow in the PR.

);

indicesRoutingToDelete.forEach(allUploadedIndicesRouting::remove);

Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

nit: avoid unnecessary line breaks

Comment on lines 125 to 148
Index index,
LatchedActionListener<IndexRoutingTable> latchedActionListener) {
int idx = uploadedFilename.lastIndexOf("/");
String blobFileName = uploadedFilename.substring(idx+1);
BlobContainer blobContainer = blobStoreRepository.blobStore().blobContainer( BlobPath.cleanPath().add(uploadedFilename.substring(0,idx)));

return () -> readAsync(
blobContainer,
blobFileName,
index,
threadPool.executor(ThreadPool.Names.GENERIC),
ActionListener.wrap(response -> latchedActionListener.onResponse(response.getIndexRoutingTable()), latchedActionListener::onFailure)
);
}

public void readAsync(BlobContainer blobContainer, String name, Index index, ExecutorService executorService, ActionListener<RemoteIndexRoutingTable> listener) throws IOException {
executorService.execute(() -> {
try {
listener.onResponse(read(blobContainer, name, index));
} catch (Exception e) {
listener.onFailure(e);
}
});
}
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Can we use interfaces from #13785

Comment on lines +832 to +838
Map<String, UploadedMetadataAttribute> customToRead,
boolean readCoordinationMetadata,
boolean readSettingsMetadata,
boolean readTransientSettingsMetadata,
boolean readTemplatesMetadata,
boolean readDiscoveryNodes,
boolean readClusterBlocks,
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Create an inner class instead?

Signed-off-by: Arpit Bandejiya <abandeji@amazon.com>
Copy link
Contributor

github-actions bot commented Jun 7, 2024

❌ Gradle check result for d254d16: null

Please examine the workflow log, locate, and copy-paste the failure(s) below, then iterate to green. Is the failure a flaky test unrelated to your change?

@Arpit-Bandejiya
Copy link
Contributor Author

Since changes are merged in #13924. Closing this PR.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

3 participants