[8.x] #101193 Preserve Step Info Across ILM Auto Retries (#113187)

docs/changelog/113187.yaml

@@ -0,0 +1,5 @@
+pr: 113187
+summary: Preserve Step Info Across ILM Auto Retries
+area: ILM+SLM
+type: enhancement
+issues: []

docs/reference/ilm/apis/explain.asciidoc

@@ -303,6 +303,12 @@ the case.
         "index_uuid": "H7lF9n36Rzqa-KfKcnGQMg",
         "index": "test-000057"
       },
+      "previous_step_info": { <5>
+        "type": "cluster_block_exception",
+        "reason": "index [test-000057/H7lF9n36Rzqa-KfKcnGQMg] blocked by: [FORBIDDEN/5/index read-only (api)",
+        "index_uuid": "H7lF9n36Rzqa-KfKcnGQMg",
+        "index": "test-000057"
+      },
       "phase_execution": {
         "policy": "my_lifecycle3",
         "phase_definition": {
@@ -329,3 +335,4 @@ is true, {ilm-init} will retry the failed step automatically.
 <3> Shows the number of attempted automatic retries to execute the failed
 step.
 <4> What went wrong
+<5> Contains a copy of the `step_info` field (when it exists) of the last attempted or executed step for diagnostic purposes, since the `step_info` is overwritten during each new attempt.

server/src/main/java/org/elasticsearch/TransportVersions.java

@@ -226,6 +226,7 @@ static TransportVersion def(int id) {
     public static final TransportVersion SEMANTIC_TEXT_SEARCH_INFERENCE_ID = def(8_750_00_0);
     public static final TransportVersion ML_INFERENCE_CHUNKING_SETTINGS = def(8_751_00_0);
     public static final TransportVersion SEMANTIC_QUERY_INNER_HITS = def(8_752_00_0);
+    public static final TransportVersion RETAIN_ILM_STEP_INFO = def(8_753_00_0);
 
     /*
      * STOP! READ THIS FIRST! No, really,

server/src/main/java/org/elasticsearch/cluster/metadata/LifecycleExecutionState.java

@@ -28,6 +28,7 @@ public record LifecycleExecutionState(
     Boolean isAutoRetryableError,
     Integer failedStepRetryCount,
     String stepInfo,
+    String previousStepInfo,
     String phaseDefinition,
     Long lifecycleDate,
     Long phaseTime,
@@ -53,6 +54,7 @@ public record LifecycleExecutionState(
     private static final String IS_AUTO_RETRYABLE_ERROR = "is_auto_retryable_error";
     private static final String FAILED_STEP_RETRY_COUNT = "failed_step_retry_count";
     private static final String STEP_INFO = "step_info";
+    private static final String PREVIOUS_STEP_INFO = "previous_step_info";
     private static final String PHASE_DEFINITION = "phase_definition";
     private static final String SNAPSHOT_NAME = "snapshot_name";
     private static final String SNAPSHOT_REPOSITORY = "snapshot_repository";
@@ -74,6 +76,7 @@ public static Builder builder(LifecycleExecutionState state) {
             .setIsAutoRetryableError(state.isAutoRetryableError)
             .setFailedStepRetryCount(state.failedStepRetryCount)
             .setStepInfo(state.stepInfo)
+            .setPreviousStepInfo(state.previousStepInfo)
             .setPhaseDefinition(state.phaseDefinition)
             .setIndexCreationDate(state.lifecycleDate)
             .setPhaseTime(state.phaseTime)
@@ -116,6 +119,10 @@ public static LifecycleExecutionState fromCustomMetadata(Map<String, String> cus
         if (stepInfo != null) {
             builder.setStepInfo(stepInfo);
         }
+        String previousStepInfo = customData.get(PREVIOUS_STEP_INFO);
+        if (previousStepInfo != null) {
+            builder.setPreviousStepInfo(previousStepInfo);
+        }
         String phaseDefinition = customData.get(PHASE_DEFINITION);
         if (phaseDefinition != null) {
             builder.setPhaseDefinition(phaseDefinition);
@@ -224,6 +231,9 @@ public Map<String, String> asMap() {
         if (stepInfo != null) {
             result.put(STEP_INFO, stepInfo);
         }
+        if (previousStepInfo != null) {
+            result.put(PREVIOUS_STEP_INFO, previousStepInfo);
+        }
         if (lifecycleDate != null) {
             result.put(INDEX_CREATION_DATE, String.valueOf(lifecycleDate));
         }
@@ -263,6 +273,7 @@ public static class Builder {
         private String step;
         private String failedStep;
         private String stepInfo;
+        private String previousStepInfo;
         private String phaseDefinition;
         private Long indexCreationDate;
         private Long phaseTime;
@@ -301,6 +312,11 @@ public Builder setStepInfo(String stepInfo) {
             return this;
         }
 
+        public Builder setPreviousStepInfo(String previousStepInfo) {
+            this.previousStepInfo = previousStepInfo;
+            return this;
+        }
+
         public Builder setPhaseDefinition(String phaseDefinition) {
             this.phaseDefinition = phaseDefinition;
             return this;
@@ -370,6 +386,7 @@ public LifecycleExecutionState build() {
                 isAutoRetryableError,
                 failedStepRetryCount,
                 stepInfo,
+                previousStepInfo,
                 phaseDefinition,
                 indexCreationDate,
                 phaseTime,

x-pack/plugin/core/src/main/java/org/elasticsearch/xpack/core/ilm/IndexLifecycleExplainResponse.java

@@ -48,6 +48,7 @@ public class IndexLifecycleExplainResponse implements ToXContentObject, Writeabl
     private static final ParseField STEP_TIME_MILLIS_FIELD = new ParseField("step_time_millis");
     private static final ParseField STEP_TIME_FIELD = new ParseField("step_time");
     private static final ParseField STEP_INFO_FIELD = new ParseField("step_info");
+    private static final ParseField PREVIOUS_STEP_INFO_FIELD = new ParseField("previous_step_info");
     private static final ParseField PHASE_EXECUTION_INFO = new ParseField("phase_execution");
     private static final ParseField AGE_FIELD = new ParseField("age");
     private static final ParseField TIME_SINCE_INDEX_CREATION_FIELD = new ParseField("time_since_index_creation");
@@ -76,6 +77,7 @@ public class IndexLifecycleExplainResponse implements ToXContentObject, Writeabl
             (String) a[17],
             (String) a[18],
             (BytesReference) a[11],
+            (BytesReference) a[21],
             (PhaseExecutionInfo) a[12]
             // a[13] == "age"
             // a[20] == "time_since_index_creation"
@@ -111,6 +113,11 @@ public class IndexLifecycleExplainResponse implements ToXContentObject, Writeabl
         PARSER.declareString(ConstructingObjectParser.optionalConstructorArg(), SHRINK_INDEX_NAME);
         PARSER.declareLong(ConstructingObjectParser.optionalConstructorArg(), INDEX_CREATION_DATE_MILLIS_FIELD);
         PARSER.declareString(ConstructingObjectParser.optionalConstructorArg(), TIME_SINCE_INDEX_CREATION_FIELD);
+        PARSER.declareObject(ConstructingObjectParser.optionalConstructorArg(), (p, c) -> {
+            XContentBuilder builder = JsonXContent.contentBuilder();
+            builder.copyCurrentStructure(p);
+            return BytesReference.bytes(builder);
+        }, PREVIOUS_STEP_INFO_FIELD);
     }
 
     private final String index;
@@ -126,6 +133,7 @@ public class IndexLifecycleExplainResponse implements ToXContentObject, Writeabl
     private final Long stepTime;
     private final boolean managedByILM;
     private final BytesReference stepInfo;
+    private final BytesReference previousStepInfo;
     private final PhaseExecutionInfo phaseExecutionInfo;
     private final Boolean isAutoRetryableError;
     private final Integer failedStepRetryCount;
@@ -153,6 +161,7 @@ public static IndexLifecycleExplainResponse newManagedIndexResponse(
         String snapshotName,
         String shrinkIndexName,
         BytesReference stepInfo,
+        BytesReference previousStepInfo,
         PhaseExecutionInfo phaseExecutionInfo
     ) {
         return new IndexLifecycleExplainResponse(
@@ -174,6 +183,7 @@ public static IndexLifecycleExplainResponse newManagedIndexResponse(
             snapshotName,
             shrinkIndexName,
             stepInfo,
+            previousStepInfo,
             phaseExecutionInfo
         );
     }
@@ -198,6 +208,7 @@ public static IndexLifecycleExplainResponse newUnmanagedIndexResponse(String ind
             null,
             null,
             null,
+            null,
             null
         );
     }
@@ -221,6 +232,7 @@ private IndexLifecycleExplainResponse(
         String snapshotName,
         String shrinkIndexName,
         BytesReference stepInfo,
+        BytesReference previousStepInfo,
         PhaseExecutionInfo phaseExecutionInfo
     ) {
         if (managedByILM) {
@@ -262,6 +274,7 @@ private IndexLifecycleExplainResponse(
                 || actionTime != null
                 || stepTime != null
                 || stepInfo != null
+                || previousStepInfo != null
                 || phaseExecutionInfo != null) {
                 throw new IllegalArgumentException(
                     "Unmanaged index response must only contain fields: [" + MANAGED_BY_ILM_FIELD + ", " + INDEX_FIELD + "]"
@@ -283,6 +296,7 @@ private IndexLifecycleExplainResponse(
         this.isAutoRetryableError = isAutoRetryableError;
         this.failedStepRetryCount = failedStepRetryCount;
         this.stepInfo = stepInfo;
+        this.previousStepInfo = previousStepInfo;
         this.phaseExecutionInfo = phaseExecutionInfo;
         this.repositoryName = repositoryName;
         this.snapshotName = snapshotName;
@@ -314,6 +328,11 @@ public IndexLifecycleExplainResponse(StreamInput in) throws IOException {
             } else {
                 indexCreationDate = null;
             }
+            if (in.getTransportVersion().onOrAfter(TransportVersions.RETAIN_ILM_STEP_INFO)) {
+                previousStepInfo = in.readOptionalBytesReference();
+            } else {
+                previousStepInfo = null;
+            }
         } else {
             policyName = null;
             lifecycleDate = null;
@@ -327,6 +346,7 @@ public IndexLifecycleExplainResponse(StreamInput in) throws IOException {
             actionTime = null;
             stepTime = null;
             stepInfo = null;
+            previousStepInfo = null;
             phaseExecutionInfo = null;
             repositoryName = null;
             snapshotName = null;
@@ -359,6 +379,9 @@ public void writeTo(StreamOutput out) throws IOException {
             if (out.getTransportVersion().onOrAfter(TransportVersions.V_8_1_0)) {
                 out.writeOptionalLong(indexCreationDate);
             }
+            if (out.getTransportVersion().onOrAfter(TransportVersions.RETAIN_ILM_STEP_INFO)) {
+                out.writeOptionalBytesReference(previousStepInfo);
+            }
         }
     }
 
@@ -422,6 +445,10 @@ public BytesReference getStepInfo() {
         return stepInfo;
     }
 
+    public BytesReference getPreviousStepInfo() {
+        return previousStepInfo;
+    }
+
     public PhaseExecutionInfo getPhaseExecutionInfo() {
         return phaseExecutionInfo;
     }
@@ -515,6 +542,9 @@ public XContentBuilder toXContent(XContentBuilder builder, Params params) throws
             if (stepInfo != null && stepInfo.length() > 0) {
                 builder.rawField(STEP_INFO_FIELD.getPreferredName(), stepInfo.streamInput(), XContentType.JSON);
             }
+            if (previousStepInfo != null && previousStepInfo.length() > 0) {
+                builder.rawField(PREVIOUS_STEP_INFO_FIELD.getPreferredName(), previousStepInfo.streamInput(), XContentType.JSON);
+            }
             if (phaseExecutionInfo != null) {
                 builder.field(PHASE_EXECUTION_INFO.getPreferredName(), phaseExecutionInfo);
             }
@@ -544,6 +574,7 @@ public int hashCode() {
             snapshotName,
             shrinkIndexName,
             stepInfo,
+            previousStepInfo,
             phaseExecutionInfo
         );
     }
@@ -575,6 +606,7 @@ public boolean equals(Object obj) {
             && Objects.equals(snapshotName, other.snapshotName)
             && Objects.equals(shrinkIndexName, other.shrinkIndexName)
             && Objects.equals(stepInfo, other.stepInfo)
+            && Objects.equals(previousStepInfo, other.previousStepInfo)
             && Objects.equals(phaseExecutionInfo, other.phaseExecutionInfo);
     }