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 2.14] fix model stuck in deploying state during node crash/cluster restart #3143

Open
wants to merge 1 commit into
base: 2.14
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 @@ -10,12 +10,16 @@
import static org.opensearch.ml.task.MLTaskManager.TASK_SEMAPHORE_TIMEOUT;
import static org.opensearch.ml.utils.MLExceptionUtils.logException;
import static org.opensearch.ml.utils.MLExceptionUtils.toJsonString;
import static org.opensearch.ml.utils.RestActionUtils.getAllNodes;

import java.time.Instant;
import java.util.Arrays;
import java.util.HashMap;
import java.util.HashSet;
import java.util.List;
import java.util.Map;
import java.util.Set;
import java.util.stream.Collectors;

import org.opensearch.action.ActionRequest;
import org.opensearch.action.support.ActionFilters;
Expand Down Expand Up @@ -131,10 +135,29 @@ protected void doExecute(Task task, ActionRequest request, ActionListener<MLForw
syncModelWorkerNodes(modelId, functionName);
}

if (workNodes == null || workNodes.size() == 0) {
Set<String> workNodesRemovedFromCluster = new HashSet<>();

if (workNodes != null && !workNodes.isEmpty()) {
Set<String> allNodesInCluster = new HashSet<>(List.of(getAllNodes(clusterService)));

workNodesRemovedFromCluster = workNodes
.stream()
.filter(node -> !allNodesInCluster.contains(node))
.collect(Collectors.toSet());

if (!workNodesRemovedFromCluster.isEmpty()) {
workNodes.removeAll(workNodesRemovedFromCluster);
}
}

if (workNodes == null || workNodes.isEmpty()) {
if (!workNodesRemovedFromCluster.isEmpty()) {
mlTaskCache.updateWorkerNode(workNodesRemovedFromCluster);
mlModelManager.removeModelWorkerNode(modelId, false, workNodesRemovedFromCluster.toArray(new String[0]));
}
int currentWorkerNodeCount = mlTaskCache.getWorkerNodeSize();
MLTaskState taskState = mlTaskCache.hasError() ? MLTaskState.COMPLETED_WITH_ERROR : MLTaskState.COMPLETED;
if (mlTaskCache.allNodeFailed()) {
if (mlTaskCache.allNodeFailed() || mlTaskCache.getWorkerNodeSize() == 0) {
taskState = MLTaskState.FAILED;
currentWorkerNodeCount = 0;
} else {
Expand All @@ -150,11 +173,11 @@ protected void doExecute(Task task, ActionRequest request, ActionListener<MLForw
mlTaskManager.updateMLTask(taskId, builder.build(), TASK_SEMAPHORE_TIMEOUT, true);

MLModelState modelState;
if (!mlTaskCache.allNodeFailed()) {
modelState = mlTaskCache.hasError() ? MLModelState.PARTIALLY_DEPLOYED : MLModelState.DEPLOYED;
} else {
if (mlTaskCache.allNodeFailed() || mlTaskCache.getWorkerNodeSize() == 0) {
modelState = MLModelState.DEPLOY_FAILED;
log.error("deploy model failed on all nodes, model id: {}", modelId);
} else {
modelState = mlTaskCache.hasError() ? MLModelState.PARTIALLY_DEPLOYED : MLModelState.DEPLOYED;
}
Map<String, Object> updateFields = new HashMap<>();
updateFields.put(MLModel.MODEL_STATE_FIELD, modelState);
Expand Down
Original file line number Diff line number Diff line change
Expand Up @@ -62,4 +62,9 @@ public int errorNodesCount() {
public boolean allNodeFailed() {
return workerNodeSize != null && errors.size() == workerNodeSize;
}

public void updateWorkerNode(Set<String> nodesRemovedFromCluster) {
this.workerNodes.removeAll(nodesRemovedFromCluster);
this.workerNodeSize = this.workerNodeSize - nodesRemovedFromCluster.size();
}
}
Original file line number Diff line number Diff line change
Expand Up @@ -29,6 +29,7 @@
import static org.opensearch.ml.settings.MLCommonsSettings.ML_COMMONS_ONLY_RUN_ON_ML_NODE;
import static org.opensearch.ml.utils.TestHelper.ML_ROLE;
import static org.opensearch.ml.utils.TestHelper.clusterSetting;
import static org.opensearch.ml.utils.TestHelper.setupTestClusterState;

import java.util.Arrays;
import java.util.HashSet;
Expand All @@ -43,6 +44,7 @@
import org.opensearch.Version;
import org.opensearch.action.support.ActionFilters;
import org.opensearch.client.Client;
import org.opensearch.cluster.ClusterState;
import org.opensearch.cluster.node.DiscoveryNode;
import org.opensearch.cluster.service.ClusterService;
import org.opensearch.common.settings.ClusterSettings;
Expand Down Expand Up @@ -94,6 +96,8 @@ public class TransportForwardActionTests extends OpenSearchTestCase {

private TransportForwardAction forwardAction;

private ClusterState testState;

Settings settings = Settings
.builder()
.put(ML_COMMONS_MODEL_AUTO_REDEPLOY_ENABLE.getKey(), true)
Expand Down Expand Up @@ -137,6 +141,9 @@ public void setup() {
)
);

testState = setupTestClusterState("test_node_id2");
when(clusterService.state()).thenReturn(testState);

node1 = new DiscoveryNode(nodeId1, buildNewFakeTransportAddress(), emptyMap(), ImmutableSet.of(ML_ROLE), Version.CURRENT);
node2 = new DiscoveryNode(nodeId2, buildNewFakeTransportAddress(), emptyMap(), ImmutableSet.of(ML_ROLE), Version.CURRENT);

Expand Down
Original file line number Diff line number Diff line change
Expand Up @@ -118,7 +118,7 @@ public void setup() throws IOException {
encryptor = spy(new EncryptorImpl(null));
syncUpCron = new MLSyncUpCron(client, clusterService, nodeHelper, mlIndicesHandler, encryptor);

testState = setupTestClusterState();
testState = setupTestClusterState("node");
when(clusterService.state()).thenReturn(testState);

doAnswer(invocation -> {
Expand Down
Original file line number Diff line number Diff line change
Expand Up @@ -151,7 +151,7 @@ public void setup() throws IOException {
.build();

clusterName = new ClusterName("test cluster");
testState = setupTestClusterState();
testState = setupTestClusterState("node");
when(clusterService.state()).thenReturn(testState);

doAnswer(invocation -> {
Expand Down
Original file line number Diff line number Diff line change
Expand Up @@ -147,7 +147,7 @@ public void setup() throws IOException {
roleSet,
Version.CURRENT
);
testState = setupTestClusterState();
testState = setupTestClusterState("node");
when(clusterService.state()).thenReturn(testState);

clusterName = new ClusterName(clusterNameStr);
Expand Down
Original file line number Diff line number Diff line change
Expand Up @@ -67,7 +67,7 @@ public class RestMLUndeployModelActionTests extends OpenSearchTestCase {
@Before
public void setup() {
MockitoAnnotations.openMocks(this);
testState = setupTestClusterState();
testState = setupTestClusterState("node");
when(clusterService.state()).thenReturn(testState);
when(clusterService.getClusterSettings()).thenReturn(clusterSettings);
restMLUndeployModelAction = new RestMLUndeployModelAction(clusterService, settings);
Expand Down
Original file line number Diff line number Diff line change
Expand Up @@ -424,11 +424,11 @@ public static ClusterState state(int numDataNodes, String indexName, String mapp
return state(new ClusterName("test"), indexName, mapping, clusterManagerNode, clusterManagerNode, allNodes);
}

public static ClusterState setupTestClusterState() {
public static ClusterState setupTestClusterState(String nodeId) {
Set<DiscoveryNodeRole> roleSet = new HashSet<>();
roleSet.add(DiscoveryNodeRole.DATA_ROLE);
DiscoveryNode node = new DiscoveryNode(
"node",
nodeId,
new TransportAddress(TransportAddress.META_ADDRESS, new AtomicInteger().incrementAndGet()),
new HashMap<>(),
roleSet,
Expand Down
Loading