Allow relocation to NOT_PREFERRED node for evacuating shards

docs/changelog/140197.yaml

@@ -0,0 +1,5 @@
+pr: 140197
+summary: Allow relocation to NOT_PREFERRED node for evacuating shards
+area: Allocation
+type: bug
+issues: []

server/src/main/java/org/elasticsearch/cluster/routing/allocation/MoveDecision.java

@@ -23,6 +23,8 @@
 import java.util.List;
 import java.util.Objects;
 
+import static org.elasticsearch.cluster.routing.allocation.decider.Decision.NO;
+
 /**
  * Represents a decision to move a started shard, either because it is no longer allowed to remain on its current node
  * or because moving it to another node will form a better cluster balance.
@@ -39,14 +41,7 @@ public final class MoveDecision extends AbstractAllocationDecision {
         null,
         0
     );
-    private static final MoveDecision CACHED_CANNOT_MOVE_DECISION = new MoveDecision(
-        null,
-        null,
-        AllocationDecision.NO,
-        Decision.NO,
-        null,
-        0
-    );
+    private static final MoveDecision CACHED_CANNOT_MOVE_DECISION = new MoveDecision(null, null, AllocationDecision.NO, NO, null, 0);
 
     @Nullable
     private final AllocationDecision canMoveDecision;
@@ -172,12 +167,11 @@ public boolean isDecisionTaken() {
      */
     public boolean cannotRemainAndCanMove() {
         checkDecisionState();
-        return cannotRemain() && (canMoveDecision == AllocationDecision.YES);
-    }
-
-    public boolean cannotRemainAndNotPreferredMove() {
-        checkDecisionState();
-        return cannotRemain() && canMoveDecision == AllocationDecision.NOT_PREFERRED;
+        return switch (canRemainDecision.type()) {
+            case NO -> canMoveDecision == AllocationDecision.YES || canMoveDecision == AllocationDecision.NOT_PREFERRED;
+            case NOT_PREFERRED -> canMoveDecision == AllocationDecision.YES;
+            default -> false;
+        };
     }
 
     /**
@@ -342,9 +336,7 @@ public Iterator<? extends ToXContent> toXContentChunked(ToXContent.Params params
                 builder.field("rebalance_explanation", getExplanation());
             } else {
                 if (cannotRemainAndCanMove()) {
-                    builder.field("can_move_to_other_node", "yes");
-                } else if (cannotRemainAndNotPreferredMove()) {
-                    builder.field("can_move_to_other_node", "not-preferred");
+                    builder.field("can_move_to_other_node", canMoveDecision);
                 } else {
                     builder.field("can_move_to_other_node", "no");
                 }
@@ -374,4 +366,17 @@ public int hashCode() {
         return 31 * super.hashCode() + Objects.hash(canMoveDecision, canRemainDecision, clusterRebalanceDecision, currentNodeRanking);
     }
 
+    @Override
+    public String toString() {
+        return "MoveDecision{"
+            + "canMoveDecision="
+            + canMoveDecision
+            + ", canRemainDecision="
+            + canRemainDecision
+            + ", clusterRebalanceDecision="
+            + clusterRebalanceDecision
+            + ", currentNodeRanking="
+            + currentNodeRanking
+            + '}';
+    }
 }

server/src/main/java/org/elasticsearch/cluster/routing/allocation/allocator/BalancedShardsAllocator.java

@@ -861,7 +861,7 @@ public boolean moveShards() {
                         shardMoved = true;
                     }
                 } else if (moveDecision.isDecisionTaken() && moveDecision.cannotRemain()) {
-                    logger.trace("[{}][{}] can't move", shardRouting.index(), shardRouting.id());
+                    logger.trace("[{}][{}] can't move: [{}]", shardRouting.index(), shardRouting.id(), moveDecision);
                 }
             }
 

x-pack/plugin/shutdown/src/internalClusterTest/java/org/elasticsearch/xpack/shutdown/ShutdownEvacuationIT.java

@@ -0,0 +1,104 @@
+/*
+ * Copyright Elasticsearch B.V. and/or licensed to Elasticsearch B.V. under one
+ * or more contributor license agreements. Licensed under the Elastic License
+ * 2.0; you may not use this file except in compliance with the Elastic License
+ * 2.0.
+ */
+
+package org.elasticsearch.xpack.shutdown;
+
+import org.elasticsearch.cluster.metadata.ProjectId;
+import org.elasticsearch.cluster.metadata.SingleNodeShutdownMetadata;
+import org.elasticsearch.cluster.routing.IndexShardRoutingTable;
+import org.elasticsearch.cluster.routing.RoutingNode;
+import org.elasticsearch.cluster.routing.ShardRouting;
+import org.elasticsearch.cluster.routing.allocation.RoutingAllocation;
+import org.elasticsearch.cluster.routing.allocation.decider.AllocationDecider;
+import org.elasticsearch.cluster.routing.allocation.decider.Decision;
+import org.elasticsearch.common.settings.ClusterSettings;
+import org.elasticsearch.common.settings.Settings;
+import org.elasticsearch.core.TimeValue;
+import org.elasticsearch.plugins.ClusterPlugin;
+import org.elasticsearch.plugins.Plugin;
+import org.elasticsearch.test.ClusterServiceUtils;
+import org.elasticsearch.test.ESIntegTestCase;
+import org.junit.After;
+
+import java.util.Arrays;
+import java.util.Collection;
+import java.util.Collections;
+import java.util.List;
+import java.util.Set;
+import java.util.concurrent.ConcurrentHashMap;
+
+import static org.elasticsearch.test.hamcrest.ElasticsearchAssertions.assertAcked;
+
+@ESIntegTestCase.ClusterScope(scope = ESIntegTestCase.Scope.TEST, numDataNodes = 0)
+public class ShutdownEvacuationIT extends ESIntegTestCase {
+
+    private static final Set<String> NOT_PREFERRED_NODES = Collections.newSetFromMap(new ConcurrentHashMap<>());
+
+    @Override
+    protected Collection<Class<? extends Plugin>> nodePlugins() {
+        return Arrays.asList(ShutdownPlugin.class, NotPreferredPlugin.class);
+    }
+
+    @After
+    public void clearNotPreferredNodes() {
+        NOT_PREFERRED_NODES.clear();
+    }
+
+    public void testCanEvacuationToNotPreferredNodeDuringShutdown() {
+        final var node1 = internalCluster().startNode();
+        final var indexName = randomIdentifier();
+        createIndex(indexName, 1, 0);
+        ensureGreen(indexName);
+        final var node2 = internalCluster().startNode();
+        final var node2ID = getNodeId(node2);
+        final var node1ID = getNodeId(node1);
+
+        NOT_PREFERRED_NODES.add(node2ID);
+
+        // Mark node 1 as shutting down
+        assertAcked(
+            internalCluster().client()
+                .execute(
+                    PutShutdownNodeAction.INSTANCE,
+                    new PutShutdownNodeAction.Request(
+                        TEST_REQUEST_TIMEOUT,
+                        TEST_REQUEST_TIMEOUT,
+                        node1ID,
+                        SingleNodeShutdownMetadata.Type.SIGTERM,
+                        "testing",
+                        null,
+                        null,
+                        TimeValue.ZERO
+                    )
+                )
+        );
+
+        safeAwait(
+            ClusterServiceUtils.addMasterTemporaryStateListener(
+                state -> state.routingTable(ProjectId.DEFAULT)
+                    .index(indexName)
+                    .allShards()
+                    .flatMap(IndexShardRoutingTable::allShards)
+                    .allMatch(shardRouting -> shardRouting.currentNodeId().equals(node2ID) && shardRouting.started())
+            )
+        );
+    }
+
+    public static class NotPreferredPlugin extends Plugin implements ClusterPlugin {
+
+        @Override
+        public Collection<AllocationDecider> createAllocationDeciders(Settings settings, ClusterSettings clusterSettings) {
+
+            return List.of(new AllocationDecider() {
+                @Override
+                public Decision canAllocate(ShardRouting shardRouting, RoutingNode node, RoutingAllocation allocation) {
+                    return NOT_PREFERRED_NODES.contains(node.nodeId()) ? Decision.NOT_PREFERRED : Decision.YES;
+                }
+            });
+        }
+    }
+}