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

[Backport 2.13] Filter out remote model auto redeployment #3215

Open
wants to merge 1 commit into
base: 2.13
Choose a base branch
from
Open
Show file tree
Hide file tree
Changes from all commits
Commits
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
Original file line number Diff line number Diff line change
Expand Up @@ -12,6 +12,7 @@
import java.util.ArrayList;
import java.util.Arrays;
import java.util.List;
import java.util.Map;
import java.util.Optional;
import java.util.Queue;
import java.util.concurrent.ConcurrentLinkedQueue;
Expand All @@ -30,6 +31,7 @@
import org.opensearch.core.common.Strings;
import org.opensearch.index.IndexNotFoundException;
import org.opensearch.index.query.TermsQueryBuilder;
import org.opensearch.ml.common.FunctionName;
import org.opensearch.ml.common.MLModel;
import org.opensearch.ml.common.model.MLModelState;
import org.opensearch.ml.common.transport.deploy.MLDeployModelAction;
Expand Down Expand Up @@ -257,17 +259,17 @@ private void queryRunningModels(ActionListener<SearchResponse> listener) {
private void triggerModelRedeploy(ModelAutoRedeployArrangement modelAutoRedeployArrangement) {
String modelId = modelAutoRedeployArrangement.getSearchResponse().getId();
List<String> addedNodes = modelAutoRedeployArrangement.getAddedNodes();
List<String> planningWorkerNodes = (List<String>) modelAutoRedeployArrangement
.getSearchResponse()
.getSourceAsMap()
.get(MLModel.PLANNING_WORKER_NODES_FIELD);
Integer autoRedeployRetryTimes = (Integer) modelAutoRedeployArrangement
.getSearchResponse()
.getSourceAsMap()
.get(MLModel.AUTO_REDEPLOY_RETRY_TIMES_FIELD);
Boolean deployToAllNodes = (Boolean) Optional
.ofNullable(modelAutoRedeployArrangement.getSearchResponse().getSourceAsMap().get(MLModel.DEPLOY_TO_ALL_NODES_FIELD))
.orElse(false);
Map<String, Object> sourceAsMap = modelAutoRedeployArrangement.getSearchResponse().getSourceAsMap();
String functionName = (String) Optional
.ofNullable(sourceAsMap.get(MLModel.FUNCTION_NAME_FIELD))
.orElse(sourceAsMap.get(MLModel.ALGORITHM_FIELD));
if (FunctionName.REMOTE == FunctionName.from(functionName)) {
log.info("Skipping redeploying remote model {} as remote model deployment can be done at prediction time.", modelId);
return;
}
List<String> planningWorkerNodes = (List<String>) sourceAsMap.get(MLModel.PLANNING_WORKER_NODES_FIELD);
Integer autoRedeployRetryTimes = (Integer) sourceAsMap.get(MLModel.AUTO_REDEPLOY_RETRY_TIMES_FIELD);
Boolean deployToAllNodes = (Boolean) Optional.ofNullable(sourceAsMap.get(MLModel.DEPLOY_TO_ALL_NODES_FIELD)).orElse(false);
// calculate node ids.
String[] nodeIds = null;
if (deployToAllNodes || !allowCustomDeploymentPlan) {
Expand Down
Original file line number Diff line number Diff line change
Expand Up @@ -609,6 +609,34 @@ public void test_redeployAModel_with_needRedeployArray_isEmpty() {
mlModelAutoReDeployer.redeployAModel();
}

public void test_buildAutoReloadArrangement_skippingRemoteModel_success() throws Exception {
Settings settings = Settings
.builder()
.put(ML_COMMONS_ONLY_RUN_ON_ML_NODE.getKey(), true)
.put(ML_COMMONS_MODEL_AUTO_REDEPLOY_LIFETIME_RETRY_TIMES.getKey(), 3)
.put(ML_COMMONS_MODEL_AUTO_REDEPLOY_ENABLE.getKey(), true)
.put(ML_COMMONS_ALLOW_CUSTOM_DEPLOYMENT_PLAN.getKey(), false)
.build();

ClusterService clusterService = mock(ClusterService.class);
when(clusterService.localNode()).thenReturn(localNode);
when(clusterService.getClusterSettings()).thenReturn(getClusterSettings(settings));
mockClusterDataNodes(clusterService);

mlModelAutoReDeployer = spy(
new MLModelAutoReDeployer(clusterService, client, settings, mlModelManager, searchRequestBuilderFactory)
);

SearchResponse searchResponse = buildDeployToAllNodesTrueSearchResponse("RemoteModelResult.json");
doAnswer(invocation -> {
ActionListener<SearchResponse> listener = invocation.getArgument(0);
listener.onResponse(searchResponse);
return null;
}).when(searchRequestBuilder).execute(isA(ActionListener.class));
mlModelAutoReDeployer.buildAutoReloadArrangement(addedNodes, clusterManagerNodeId);
verify(client, never()).execute(any(MLDeployModelAction.class), any(MLDeployModelRequest.class), any(ActionListener.class));
}

private SearchResponse buildDeployToAllNodesTrueSearchResponse(String file) throws Exception {
MLModel mlModel = buildModelWithJsonFile(file);
return createResponseWithModel(mlModel);
Expand Down
Original file line number Diff line number Diff line change
@@ -0,0 +1,20 @@
{
"last_deployed_time": 1722954415807,
"model_version": "619",
"created_time": 1722954415642,
"deploy_to_all_nodes": true,
"is_hidden": false,
"description": "This is a test model",
"model_state": "DEPLOYED",
"planning_worker_node_count": 1,
"auto_redeploy_retry_times": 0,
"last_updated_time": 1723691017054,
"name": "my sagemaker model",
"connector_id": "z3kVKJEBAfFjoGUT_Ui7",
"current_worker_node_count": 0,
"model_group_id": "MiJPJ5EBQM-QzppeWrTJ",
"planning_worker_nodes": [
"DecGG5pDQYaqelLMLcIV9Q"
],
"algorithm": "REMOTE"
}
Loading