Remove snapshotpath, index type from snapshots

slackhq · Aug 14, 2024 · 89b7d25 · 89b7d25
1 parent a50d63e
commit 89b7d25
Show file tree

Hide file tree

Showing 42 changed files with 272 additions and 1,395 deletions.
diff --git a/astra/src/main/java/com/slack/astra/chunk/ChunkInfo.java b/astra/src/main/java/com/slack/astra/chunk/ChunkInfo.java
@@ -3,7 +3,6 @@
 import static com.slack.astra.util.ArgValidationUtils.ensureTrue;
 
 import com.slack.astra.metadata.snapshot.SnapshotMetadata;
-import com.slack.astra.proto.metadata.Metadata;
 import java.util.Objects;
 
 /**
@@ -31,19 +30,16 @@ public static ChunkInfo fromSnapshotMetadata(SnapshotMetadata snapshotMetadata)
         snapshotMetadata.endTimeEpochMs,
         snapshotMetadata.maxOffset,
         snapshotMetadata.partitionId,
-        snapshotMetadata.snapshotPath,
         snapshotMetadata.sizeInBytesOnDisk);
   }
 
   public static SnapshotMetadata toSnapshotMetadata(ChunkInfo chunkInfo, String chunkPrefix) {
     return new SnapshotMetadata(
         chunkPrefix + chunkInfo.chunkId,
-        chunkInfo.snapshotPath,
         chunkInfo.getDataStartTimeEpochMs(),
         chunkInfo.getDataEndTimeEpochMs(),
         chunkInfo.maxOffset,
         chunkInfo.kafkaPartitionId,
-        Metadata.IndexType.LOGS_LUCENE9,
         chunkInfo.sizeInBytesOnDisk);
   }
 
@@ -77,14 +73,10 @@ public static SnapshotMetadata toSnapshotMetadata(ChunkInfo chunkInfo, String ch
   // indexing and snapshotting and is not useful afterwards.
   private long chunkSnapshotTimeEpochMs;
 
-  // Path to S3 snapshot.
-  private String snapshotPath;
-
   // Size of chunk on disk in bytes
   private long sizeInBytesOnDisk;
 
-  public ChunkInfo(
-      String chunkId, long chunkCreationTimeEpochMs, String kafkaPartitionId, String snapshotPath) {
+  public ChunkInfo(String chunkId, long chunkCreationTimeEpochMs, String kafkaPartitionId) {
     // TODO: Should we set the snapshot time to creation time also?
     this(
         chunkId,
@@ -95,7 +87,6 @@ public ChunkInfo(
         0,
         DEFAULT_MAX_OFFSET,
         kafkaPartitionId,
-        snapshotPath,
         0);
   }
 
@@ -108,7 +99,6 @@ public ChunkInfo(
       long chunkSnapshotTimeEpochMs,
       long maxOffset,
       String kafkaPartitionId,
-      String snapshotPath,
       long sizeInBytesOnDisk) {
     ensureTrue(chunkId != null && !chunkId.isEmpty(), "Invalid chunk dataset name " + chunkId);
     ensureTrue(
@@ -124,7 +114,6 @@ public ChunkInfo(
     this.chunkSnapshotTimeEpochMs = chunkSnapshotTimeEpochMs;
     this.maxOffset = maxOffset;
     this.kafkaPartitionId = kafkaPartitionId;
-    this.snapshotPath = snapshotPath;
     this.sizeInBytesOnDisk = sizeInBytesOnDisk;
   }
 
@@ -164,14 +153,6 @@ public void setChunkLastUpdatedTimeEpochMs(long chunkLastUpdatedTimeEpochMs) {
     this.chunkLastUpdatedTimeEpochMs = chunkLastUpdatedTimeEpochMs;
   }
 
-  public void setSnapshotPath(String snapshotPath) {
-    this.snapshotPath = snapshotPath;
-  }
-
-  public String getSnapshotPath() {
-    return snapshotPath;
-  }
-
   public long getSizeInBytesOnDisk() {
     return sizeInBytesOnDisk;
   }
@@ -238,8 +219,6 @@ public String toString() {
         + dataEndTimeEpochMs
         + ", chunkSnapshotTimeEpochMs="
         + chunkSnapshotTimeEpochMs
-        + ", snapshotPath='"
-        + snapshotPath
         + ", sizeInBytesOnDisk='"
         + sizeInBytesOnDisk
         + '}';
@@ -258,7 +237,6 @@ public boolean equals(Object o) {
         && chunkSnapshotTimeEpochMs == chunkInfo.chunkSnapshotTimeEpochMs
         && Objects.equals(chunkId, chunkInfo.chunkId)
         && Objects.equals(kafkaPartitionId, chunkInfo.kafkaPartitionId)
-        && Objects.equals(snapshotPath, chunkInfo.snapshotPath)
         && sizeInBytesOnDisk == chunkInfo.sizeInBytesOnDisk;
   }
 
@@ -272,7 +250,6 @@ public int hashCode() {
         chunkLastUpdatedTimeEpochMs,
         dataStartTimeEpochMs,
         dataEndTimeEpochMs,
-        chunkSnapshotTimeEpochMs,
-        snapshotPath);
+        chunkSnapshotTimeEpochMs);
   }
 }
diff --git a/astra/src/main/java/com/slack/astra/chunk/IndexingChunkImpl.java b/astra/src/main/java/com/slack/astra/chunk/IndexingChunkImpl.java
@@ -6,7 +6,6 @@
 import com.slack.astra.metadata.search.SearchMetadataStore;
 import com.slack.astra.metadata.snapshot.SnapshotMetadata;
 import com.slack.astra.metadata.snapshot.SnapshotMetadataStore;
-import com.slack.astra.proto.metadata.Metadata;
 import io.micrometer.core.instrument.MeterRegistry;
 import org.slf4j.Logger;
 import org.slf4j.LoggerFactory;
@@ -66,18 +65,16 @@ public void postSnapshot() {
     SnapshotMetadata nonLiveSnapshotMetadata = toSnapshotMetadata(chunkInfo, "");
     snapshotMetadataStore.createSync(nonLiveSnapshotMetadata);
 
-    // Update the live snapshot. Keep the same snapshotId and snapshotPath to
+    // Update the live snapshot. Keep the same snapshotId to
     // ensure it's a live snapshot.
     SnapshotMetadata updatedSnapshotMetadata =
         new SnapshotMetadata(
             liveSnapshotMetadata.snapshotId,
-            liveSnapshotMetadata.snapshotPath,
             chunkInfo.getDataStartTimeEpochMs(),
             chunkInfo.getDataEndTimeEpochMs(),
             chunkInfo.getMaxOffset(),
             chunkInfo.getKafkaPartitionId(),
-            Metadata.IndexType.LOGS_LUCENE9,
-            chunkInfo.getSizeInBytesOnDisk());
+            liveSnapshotMetadata.sizeInBytesOnDisk);
     snapshotMetadataStore.updateSync(updatedSnapshotMetadata);
     liveSnapshotMetadata = updatedSnapshotMetadata;
 

diff --git a/astra/src/main/java/com/slack/astra/chunk/ReadOnlyChunkImpl.java b/astra/src/main/java/com/slack/astra/chunk/ReadOnlyChunkImpl.java
@@ -29,7 +29,6 @@
 import java.nio.file.Files;
 import java.nio.file.Path;
 import java.time.Instant;
-import java.util.List;
 import java.util.Map;
 import java.util.Objects;
 import java.util.UUID;
@@ -152,7 +151,6 @@ public ReadOnlyChunkImpl(
               Metadata.CacheSlotMetadata.CacheSlotState.FREE,
               "",
               Instant.now().toEpochMilli(),
-              List.of(Metadata.IndexType.LOGS_LUCENE9),
               searchContext.hostname,
               replicaSet);
       cacheSlotMetadataStore.createSync(cacheSlotMetadata);
@@ -265,9 +263,8 @@ public void downloadChunkData() {
       long durationNanos = assignmentTimer.stop(chunkAssignmentTimerSuccess);
 
       LOG.info(
-          "Downloaded chunk with snapshot id '{}' at path '{}' in {} seconds, was {}",
+          "Downloaded chunk with snapshot id '{}' in {} seconds, was {}",
           snapshotMetadata.snapshotId,
-          snapshotMetadata.snapshotPath,
           TimeUnit.SECONDS.convert(durationNanos, TimeUnit.NANOSECONDS),
           FileUtils.byteCountToDisplaySize(FileUtils.sizeOfDirectory(dataDirectory.toFile())));
     } catch (Exception e) {
@@ -408,9 +405,8 @@ private void handleChunkAssignment(CacheSlotMetadata cacheSlotMetadata) {
       long durationNanos = assignmentTimer.stop(chunkAssignmentTimerSuccess);
 
       LOG.debug(
-          "Downloaded chunk with snapshot id '{}' at path '{}' in {} seconds, was {}",
+          "Downloaded chunk with snapshot id '{}' in {} seconds, was {}",
           snapshotMetadata.snapshotId,
-          snapshotMetadata.snapshotPath,
           TimeUnit.SECONDS.convert(durationNanos, TimeUnit.NANOSECONDS),
           FileUtils.byteCountToDisplaySize(FileUtils.sizeOfDirectory(dataDirectory.toFile())));
     } catch (Exception e) {

diff --git a/astra/src/main/java/com/slack/astra/chunk/ReadWriteChunk.java b/astra/src/main/java/com/slack/astra/chunk/ReadWriteChunk.java
@@ -68,7 +68,7 @@ public abstract class ReadWriteChunk<T> implements Chunk<T> {
   public static final String INDEX_FILES_UPLOAD = "index_files_upload";
   public static final String INDEX_FILES_UPLOAD_FAILED = "index_files_upload_failed";
   public static final String SNAPSHOT_TIMER = "snapshot.timer";
-  public static final String LIVE_SNAPSHOT_PREFIX = SnapshotMetadata.LIVE_SNAPSHOT_PATH + "_";
+  public static final String LIVE_SNAPSHOT_PREFIX = "LIVE_";
   public static final String SCHEMA_FILE_NAME = "schema.json";
 
   private final LogStore logStore;
@@ -111,8 +111,7 @@ protected ReadWriteChunk(
         new ChunkInfo(
             chunkDataPrefix + "_" + chunkCreationTime.getEpochSecond() + "_" + logStoreId,
             chunkCreationTime.toEpochMilli(),
-            kafkaPartitionId,
-            SnapshotMetadata.LIVE_SNAPSHOT_PATH);
+            kafkaPartitionId);
 
     readOnly = false;
     this.meterRegistry = meterRegistry;
@@ -253,7 +252,6 @@ public boolean snapshotToS3(ChunkStore chunkStore) {
       chunkStore.upload(chunkInfo.chunkId, dirPath);
 
       snapshotTimer.stop(meterRegistry.timer(SNAPSHOT_TIMER));
-      chunkInfo.setSnapshotPath(chunkStore.getRemotePath(chunkInfo.chunkId));
       chunkInfo.setSizeInBytesOnDisk(totalBytes);
       logger.info("Finished RW chunk snapshot to S3 {}.", chunkInfo);
       return true;

diff --git a/astra/src/main/java/com/slack/astra/chunkManager/ChunkManagerBase.java b/astra/src/main/java/com/slack/astra/chunkManager/ChunkManagerBase.java
@@ -98,7 +98,6 @@ public SearchResult<T> query(SearchQuery query, Duration queryTimeout) {
                                       Tracing.currentTracer()
                                           .startScopedSpan("ChunkManagerBase.chunkQuery");
                                   span.tag("chunkId", chunk.id());
-                                  span.tag("chunkSnapshotPath", chunk.info().getSnapshotPath());
                                   concurrentQueries.acquire();
                                   try {
                                     return chunk.query(query);

diff --git a/astra/src/main/java/com/slack/astra/clusterManager/ReplicaCreationService.java b/astra/src/main/java/com/slack/astra/clusterManager/ReplicaCreationService.java
@@ -18,7 +18,6 @@
 import com.slack.astra.metadata.snapshot.SnapshotMetadata;
 import com.slack.astra.metadata.snapshot.SnapshotMetadataStore;
 import com.slack.astra.proto.config.AstraConfigs;
-import com.slack.astra.proto.metadata.Metadata;
 import io.micrometer.core.instrument.Counter;
 import io.micrometer.core.instrument.MeterRegistry;
 import io.micrometer.core.instrument.Timer;
@@ -177,7 +176,7 @@ protected Map<String, Integer> createReplicasForUnassignedSnapshots() {
               .filter(
                   snapshotMetadata ->
                       snapshotMetadata.endTimeEpochMs > snapshotExpiration
-                          && !SnapshotMetadata.isLive(snapshotMetadata)
+                          && !snapshotMetadata.isLive()
                           && !existingReplicas.contains(snapshotMetadata.snapshotId))
               .map(
                   (snapshotMetadata) -> {
@@ -249,7 +248,6 @@ public static ReplicaMetadata replicaMetadataFromSnapshotId(
         replicaSet,
         Instant.now().toEpochMilli(),
         expireAfter.toEpochMilli(),
-        isRestored,
-        Metadata.IndexType.LOGS_LUCENE9);
+        isRestored);
   }
 }
diff --git a/astra/src/main/java/com/slack/astra/clusterManager/SnapshotDeletionService.java b/astra/src/main/java/com/slack/astra/clusterManager/SnapshotDeletionService.java
@@ -14,7 +14,6 @@
 import com.google.common.util.concurrent.ThreadFactoryBuilder;
 import com.slack.astra.blobfs.ChunkStore;
 import com.slack.astra.metadata.replica.ReplicaMetadataStore;
-import com.slack.astra.metadata.snapshot.SnapshotMetadata;
 import com.slack.astra.metadata.snapshot.SnapshotMetadataStore;
 import com.slack.astra.proto.config.AstraConfigs;
 import io.micrometer.core.instrument.Counter;
@@ -175,7 +174,7 @@ protected int deleteExpiredSnapshotsWithoutReplicas() {
             // served from the indexers. To avoid the whole headache of managing all the
             // different states we could be in, we should just disable the deletion of live
             // snapshots whole-cloth. We clean those up when a node boots anyhow
-            .filter(snapshotMetadata -> !SnapshotMetadata.isLive(snapshotMetadata))
+            .filter(snapshotMetadata -> !snapshotMetadata.isLive())
             .map(
                 snapshotMetadata -> {
                   ListenableFuture<?> future =

diff --git a/astra/src/main/java/com/slack/astra/metadata/cache/CacheSlotMetadata.java b/astra/src/main/java/com/slack/astra/metadata/cache/CacheSlotMetadata.java
@@ -4,8 +4,6 @@
 
 import com.slack.astra.metadata.core.AstraPartitionedMetadata;
 import com.slack.astra.proto.metadata.Metadata;
-import java.util.Collections;
-import java.util.List;
 import java.util.Objects;
 
 /**
@@ -18,23 +16,18 @@ public class CacheSlotMetadata extends AstraPartitionedMetadata {
   public final Metadata.CacheSlotMetadata.CacheSlotState cacheSlotState;
   public final String replicaId;
   public final long updatedTimeEpochMs;
-  public final List<Metadata.IndexType> supportedIndexTypes;
 
   public CacheSlotMetadata(
       String name,
       Metadata.CacheSlotMetadata.CacheSlotState cacheSlotState,
       String replicaId,
       long updatedTimeEpochMs,
-      List<Metadata.IndexType> supportedIndexTypes,
       String hostname,
       String replicaSet) {
     super(name);
     checkArgument(hostname != null && !hostname.isEmpty(), "Hostname cannot be null or empty");
     checkArgument(cacheSlotState != null, "Cache slot state cannot be null");
     checkArgument(updatedTimeEpochMs > 0, "Updated time must be greater than 0");
-    checkArgument(
-        supportedIndexTypes != null && !supportedIndexTypes.isEmpty(),
-        "supported index types shouldn't be empty");
     if (cacheSlotState.equals(Metadata.CacheSlotMetadata.CacheSlotState.FREE)) {
       checkArgument(
           replicaId != null && replicaId.isEmpty(),
@@ -50,7 +43,6 @@ public CacheSlotMetadata(
     this.cacheSlotState = cacheSlotState;
     this.replicaId = replicaId;
     this.updatedTimeEpochMs = updatedTimeEpochMs;
-    this.supportedIndexTypes = Collections.unmodifiableList(supportedIndexTypes);
   }
 
   public String getHostname() {
@@ -67,8 +59,7 @@ public boolean equals(Object o) {
     if (!hostname.equals(that.hostname)) return false;
     if (!Objects.equals(replicaSet, that.replicaSet)) return false;
     if (cacheSlotState != that.cacheSlotState) return false;
-    if (!replicaId.equals(that.replicaId)) return false;
-    return supportedIndexTypes.equals(that.supportedIndexTypes);
+    return replicaId.equals(that.replicaId);
   }
 
   @Override
@@ -79,7 +70,6 @@ public int hashCode() {
     result = 31 * result + cacheSlotState.hashCode();
     result = 31 * result + replicaId.hashCode();
     result = 31 * result + (int) (updatedTimeEpochMs ^ (updatedTimeEpochMs >>> 32));
-    result = 31 * result + supportedIndexTypes.hashCode();
     return result;
   }
 
@@ -99,8 +89,6 @@ public String toString() {
         + '\''
         + ", updatedTimeEpochMs="
         + updatedTimeEpochMs
-        + ", supportedIndexTypes="
-        + supportedIndexTypes
         + ", name='"
         + name
         + '\''

diff --git a/astra/src/main/java/com/slack/astra/metadata/cache/CacheSlotMetadataSerializer.java b/astra/src/main/java/com/slack/astra/metadata/cache/CacheSlotMetadataSerializer.java
@@ -14,7 +14,6 @@ private static CacheSlotMetadata fromCacheSlotMetadataProto(
             cacheSlotMetadataProto.getCacheSlotState().name()),
         cacheSlotMetadataProto.getReplicaId(),
         cacheSlotMetadataProto.getUpdatedTimeEpochMs(),
-        cacheSlotMetadataProto.getSupportedIndexTypesList(),
         cacheSlotMetadataProto.getHostname(),
         cacheSlotMetadataProto.getReplicaSet());
   }
@@ -25,7 +24,6 @@ private static Metadata.CacheSlotMetadata toCacheSlotMetadataProto(CacheSlotMeta
         .setReplicaId(metadata.replicaId)
         .setCacheSlotState(metadata.cacheSlotState)
         .setUpdatedTimeEpochMs(metadata.updatedTimeEpochMs)
-        .addAllSupportedIndexTypes(metadata.supportedIndexTypes)
         .setHostname(metadata.hostname)
         .setReplicaSet(metadata.replicaSet)
         .build();

diff --git a/astra/src/main/java/com/slack/astra/metadata/cache/CacheSlotMetadataStore.java b/astra/src/main/java/com/slack/astra/metadata/cache/CacheSlotMetadataStore.java
@@ -50,7 +50,6 @@ public ListenableFuture<?> updateCacheSlotStateStateWithReplicaId(
             newState,
             replicaId,
             Instant.now().toEpochMilli(),
-            cacheSlotMetadata.supportedIndexTypes,
             cacheSlotMetadata.hostname,
             cacheSlotMetadata.replicaSet);
     // todo - consider refactoring this to return a completable future instead