ibessonov commented on code in PR #4769:
URL: https://github.com/apache/ignite-3/pull/4769#discussion_r1876024833


##########
modules/distribution-zones/src/main/java/org/apache/ignite/internal/distributionzones/rebalance/RebalanceMinimumRequiredTimeProviderImpl.java:
##########
@@ -0,0 +1,294 @@
+/*
+ * Licensed to the Apache Software Foundation (ASF) under one or more
+ * contributor license agreements. See the NOTICE file distributed with
+ * this work for additional information regarding copyright ownership.
+ * The ASF licenses this file to You under the Apache License, Version 2.0
+ * (the "License"); you may not use this file except in compliance with
+ * the License. You may obtain a copy of the License at
+ *
+ *      http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+package org.apache.ignite.internal.distributionzones.rebalance;
+
+import static java.lang.Math.min;
+import static java.util.Collections.emptyMap;
+import static java.util.stream.Collectors.toMap;
+import static 
org.apache.ignite.internal.distributionzones.rebalance.RebalanceUtil.PENDING_ASSIGNMENTS_PREFIX_BYTES;
+import static 
org.apache.ignite.internal.distributionzones.rebalance.RebalanceUtil.PENDING_CHANGE_TRIGGER_PREFIX_BYTES;
+import static 
org.apache.ignite.internal.distributionzones.rebalance.RebalanceUtil.STABLE_ASSIGNMENTS_PREFIX_BYTES;
+
+import java.util.HashMap;
+import java.util.HashSet;
+import java.util.Iterator;
+import java.util.Map;
+import java.util.NavigableMap;
+import java.util.Set;
+import java.util.TreeMap;
+import org.apache.ignite.internal.catalog.Catalog;
+import org.apache.ignite.internal.catalog.CatalogService;
+import org.apache.ignite.internal.catalog.descriptors.CatalogTableDescriptor;
+import org.apache.ignite.internal.catalog.descriptors.CatalogZoneDescriptor;
+import org.apache.ignite.internal.lang.ByteArray;
+import org.apache.ignite.internal.metastorage.Entry;
+import org.apache.ignite.internal.metastorage.MetaStorageManager;
+import org.apache.ignite.internal.partitiondistribution.Assignments;
+import org.apache.ignite.internal.replicator.TablePartitionId;
+import org.apache.ignite.internal.util.ByteUtils;
+import org.apache.ignite.internal.util.Cursor;
+
+/**
+ * {@link RebalanceMinimumRequiredTimeProvider} implementation for the current 
implementation of assignments. Assumes that each table has
+ * its own assignments, but assignments within a zone are still somehow 
coordinated.
+ */
+public class RebalanceMinimumRequiredTimeProviderImpl implements 
RebalanceMinimumRequiredTimeProvider {
+    private final MetaStorageManager metaStorageManager;
+    private final CatalogService catalogService;
+
+    /**
+     * Constructor.
+     */
+    public RebalanceMinimumRequiredTimeProviderImpl(MetaStorageManager 
metaStorageManager, CatalogService catalogService) {
+        this.metaStorageManager = metaStorageManager;
+        this.catalogService = catalogService;
+    }
+
+    @Override
+    public long minimumRequiredTime() {
+        // Use the same revision to read all the data, in order to guarantee 
consistency of data.
+        long appliedRevision = metaStorageManager.appliedRevision();
+
+        // Ignore the real safe time, having a time associated with revision 
is enough. Also, acquiring real safe time would be
+        // unnecessarily more complicated due to handling of possible data 
races.
+        long metaStorageSafeTime = 
metaStorageManager.timestampByRevisionLocally(appliedRevision).longValue();
+
+        long minTimestamp = metaStorageSafeTime;
+
+        Map<Integer, Map<Integer, Assignments>> stableAssignments = 
readAssignments(STABLE_ASSIGNMENTS_PREFIX_BYTES, appliedRevision);
+        Map<Integer, Map<Integer, Assignments>> pendingAssignments = 
readAssignments(PENDING_ASSIGNMENTS_PREFIX_BYTES, appliedRevision);
+
+        Map<Integer, Long> pendingChangeTriggerRevisions = 
readPendingChangeTriggerRevisions(
+                PENDING_CHANGE_TRIGGER_PREFIX_BYTES,
+                appliedRevision
+        );
+
+        int earliestCatalogVersion = catalogService.earliestCatalogVersion();
+        int latestCatalogVersion = catalogService.latestCatalogVersion();
+
+        Map<Integer, Integer> tableIdToZoneIdMap = 
tableIdToZoneIdMap(earliestCatalogVersion, latestCatalogVersion);
+        Map<Integer, NavigableMap<Long, CatalogZoneDescriptor>> 
allZonesByTimestamp = allZonesByTimestamp(
+                earliestCatalogVersion,
+                latestCatalogVersion
+        );
+        Map<Integer, NavigableMap<Long, CatalogZoneDescriptor>> 
allZonesByRevision = allZonesByRevision(allZonesByTimestamp);
+        Map<Integer, Long> zoneDeletionTimestamps = 
zoneDeletionTimestamps(earliestCatalogVersion, latestCatalogVersion);
+
+        for (Map.Entry<Integer, Integer> entry : 
tableIdToZoneIdMap.entrySet()) {
+            Integer tableId = entry.getKey();
+            Integer zoneId = entry.getValue();
+
+            NavigableMap<Long, CatalogZoneDescriptor> zoneDescriptors = 
allZonesByTimestamp.get(zoneId);
+            int zonePartitions = 
zoneDescriptors.lastEntry().getValue().partitions();
+
+            Long pendingChangeTriggerKey = 
pendingChangeTriggerRevisions.get(tableId);
+
+            // +-1 here ir required for 2 reasons:
+            // - we need timestamp right before deletion, if zone is deleted, 
thus we must subtract 1;
+            // - we need a "metaStorageSafeTime" if zone is not deleted, 
without any subtractions.
+            long latestTimestamp = zoneDeletionTimestamps.getOrDefault(zoneId, 
metaStorageSafeTime + 1) - 1;
+
+            long zoneRevision = pendingChangeTriggerKey == null
+                    ? zoneDescriptors.firstEntry().getValue().updateToken()
+                    : pendingChangeTriggerKey;
+
+            NavigableMap<Long, CatalogZoneDescriptor> map = 
allZonesByRevision.get(zoneId);
+            Map.Entry<Long, CatalogZoneDescriptor> zone = 
map.floorEntry(zoneRevision);
+            long timestamp = 
metaStorageManager.timestampByRevisionLocally(zone.getValue().updateToken()).longValue();
+
+            timestamp = ceilTime(zoneDescriptors, timestamp, latestTimestamp);

Review Comment:
   Updated the code according to this comment. Now I distinguish between zone 
updates that do lead or don't lead to reassignment.



-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: notifications-unsubscr...@ignite.apache.org

For queries about this service, please contact Infrastructure at:
us...@infra.apache.org

Reply via email to