| /* |
| * 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.geode.internal.cache; |
| |
| import static java.util.concurrent.TimeUnit.MINUTES; |
| import static org.apache.geode.internal.cache.PartitionedRegion.RETRY_TIMEOUT_PROPERTY; |
| import static org.apache.geode.test.dunit.Host.getHost; |
| import static org.apache.geode.test.dunit.IgnoredException.addIgnoredException; |
| import static org.assertj.core.api.Assertions.assertThat; |
| |
| import java.util.NoSuchElementException; |
| import java.util.Set; |
| import java.util.concurrent.CountDownLatch; |
| |
| import org.junit.Before; |
| import org.junit.Rule; |
| import org.junit.Test; |
| |
| import org.apache.geode.cache.Cache; |
| import org.apache.geode.cache.DataPolicy; |
| import org.apache.geode.cache.PartitionAttributesFactory; |
| import org.apache.geode.cache.PartitionedRegionStorageException; |
| import org.apache.geode.cache.Region; |
| import org.apache.geode.cache.RegionFactory; |
| import org.apache.geode.cache.RegionShortcut; |
| import org.apache.geode.distributed.internal.membership.InternalDistributedMember; |
| import org.apache.geode.internal.cache.control.InternalResourceManager; |
| import org.apache.geode.internal.cache.control.InternalResourceManager.ResourceObserver; |
| import org.apache.geode.internal.cache.control.InternalResourceManager.ResourceObserverAdapter; |
| import org.apache.geode.internal.cache.partitioned.RegionAdvisor; |
| import org.apache.geode.test.dunit.IgnoredException; |
| import org.apache.geode.test.dunit.VM; |
| import org.apache.geode.test.dunit.cache.CacheTestCase; |
| import org.apache.geode.test.dunit.rules.DistributedRestoreSystemProperties; |
| |
| |
| public class PartitionedRegionHADUnitTest extends CacheTestCase { |
| |
| private static final String PR_ZERO_REDUNDANCY = "PR_ZeroRedundancy"; |
| private static final String PR_ONE_REDUNDANCY = "PR_SingleRedundancy"; |
| private static final String RETRY_TIMEOUT_VALUE = "20000"; |
| |
| private String regionName; |
| private String prNamePrefix; |
| private int numberOfRegions; |
| private int totalNumBuckets; |
| private int redundantCopies; |
| |
| private VM vm0; |
| private VM vm1; |
| private VM vm2; |
| private VM vm3; |
| |
| private VM datastoreVM0; |
| private VM datastoreVM2; |
| private VM accessorVM3; |
| |
| @Rule |
| public DistributedRestoreSystemProperties restoreSystemProperties = |
| new DistributedRestoreSystemProperties(); |
| |
| @Before |
| public void setUp() throws Exception { |
| regionName = getUniqueName(); |
| prNamePrefix = "PR-"; |
| numberOfRegions = 1; |
| totalNumBuckets = 5; |
| redundantCopies = 0; |
| |
| vm0 = getHost(0).getVM(0); |
| vm1 = getHost(0).getVM(1); |
| vm2 = getHost(0).getVM(2); |
| vm3 = getHost(0).getVM(3); |
| |
| datastoreVM0 = vm0; |
| datastoreVM2 = vm2; |
| accessorVM3 = vm3; |
| } |
| |
| /** |
| * Test to ensure that we have proper bucket failover, with no data loss, in the face of |
| * sequential cache.close() events. |
| */ |
| @Test |
| public void testBucketFailOverDuringCacheClose() throws Exception { |
| Boolean value = Boolean.TRUE; |
| |
| vm2.invoke(() -> createPartitionedRegion()); |
| vm3.invoke(() -> createPartitionedRegion()); |
| |
| vm3.invoke(() -> { |
| Cache cache = getCache(); |
| PartitionedRegion partitionedRegion = (PartitionedRegion) cache.getRegion(regionName); |
| assertThat(partitionedRegion).isEmpty(); |
| |
| // Create keys such that all buckets are created, Integer works well |
| // assuming buckets are allocated on the mod of the key hashCode, x 2 just to be safe |
| int numEntries = partitionedRegion.getTotalNumberOfBuckets() * 2; |
| |
| for (int i = numEntries; i >= 0; --i) { |
| partitionedRegion.put(i, value); |
| } |
| |
| assertThat(partitionedRegion).hasSize(numEntries + 1); |
| assertThat(partitionedRegion.getRegionAdvisor().getBucketSet()) |
| .hasSize(partitionedRegion.getTotalNumberOfBuckets()); |
| }); |
| |
| vm3.invoke(() -> validateEntries(value)); |
| vm2.invoke(() -> validateEntries(value)); |
| |
| // origin VM down! |
| vm2.invoke(() -> getCache().close()); |
| // origin down, but no data loss |
| vm3.invoke(() -> validateEntries(value)); |
| |
| // get back to the desired redundancy |
| vm0.invoke(() -> createPartitionedRegion()); |
| // verify no data loss |
| vm0.invoke(() -> validateEntries(value)); |
| |
| // 2nd oldest VM down! |
| vm3.invoke(() -> getCache().close()); |
| // 2nd down, but no data loss |
| vm0.invoke(() -> validateEntries(value)); |
| |
| // get back (for 2nd time) to desired redundancy |
| vm1.invoke(() -> createPartitionedRegion()); |
| // verify no data loss |
| vm1.invoke(() -> validateEntries(value)); |
| vm0.invoke(() -> validateEntries(value)); |
| } |
| |
| @Test |
| public void testGrabBackupBuckets() throws Exception { |
| redundantCopies = 1; |
| |
| datastoreVM0.invoke(() -> createPRsAndAwaitRecovery(200, redundantCopies, totalNumBuckets)); |
| |
| // Do put operations on these 2 PRs asynchronously (this test does not currently do this) |
| |
| try (IgnoredException ie = addIgnoredException(PartitionedRegionStorageException.class)) { |
| datastoreVM0.invoke(this::putsInDatastoreVM0); |
| } |
| |
| // At this point redundancy criterion is not meet. |
| // now if we create PRs on more VMs, it should create those "supposed to |
| // be redundant" buckets on these nodes, if it can accommodate the data |
| // (localMaxMemory>0). |
| datastoreVM2.invoke(() -> createPRsAndAwaitRecovery(200, redundantCopies, totalNumBuckets)); |
| |
| datastoreVM0.invoke(this::putsInDatastoreVM0); |
| |
| accessorVM3.invoke(() -> { |
| Cache cache = getCache(); |
| for (int i = 0; i < numberOfRegions; i++) { |
| String regionName = prNamePrefix + i; |
| createPartitionedRegion(cache, regionName, 0, redundantCopies, totalNumBuckets); |
| } |
| }); |
| |
| for (int i = 0; i < numberOfRegions; i++) { |
| final int whichRegion = i; |
| |
| int vm2LBRsize = |
| (Integer) datastoreVM2.invoke(() -> validateLocalBucket2RegionMapSize(whichRegion)); |
| int vm3LBRsize = |
| (Integer) accessorVM3.invoke(() -> validateLocalBucket2RegionMapSize(whichRegion)); |
| |
| // This would mean that up coming node didn't pick up any buckets |
| assertThat(vm2LBRsize).isGreaterThan(0); |
| |
| // This accessor should NOT have picked up any buckets. |
| assertThat(vm3LBRsize).isEqualTo(0); |
| |
| int vm2B2Nsize = (Integer) datastoreVM2.invoke(() -> validateBucketsOnNode(whichRegion)); |
| assertThat(vm2LBRsize).isEqualTo(vm2B2Nsize); |
| } |
| } |
| |
| /** |
| * This verifies the Bucket Regions on the basis of redundantCopies set in RegionAttributes. |
| */ |
| @Test |
| public void testBucketsScope() throws Exception { |
| // Create PRs on only 2 VMs |
| vm0.invoke(() -> createPRs(PR_ZERO_REDUNDANCY, PR_ONE_REDUNDANCY)); |
| vm1.invoke(() -> createPRs(PR_ZERO_REDUNDANCY, PR_ONE_REDUNDANCY)); |
| |
| // Do put operations on these 2 PRs asynchronously (test does not currently do this) |
| |
| vm0.invoke(() -> { |
| Cache cache = getCache(); |
| Region<Integer, Integer> regionZeroRedundancy = cache.getRegion(PR_ZERO_REDUNDANCY); |
| for (int k = 0; k < 10; k++) { |
| regionZeroRedundancy.put(k, k); |
| } |
| |
| Region<Integer, Integer> regionOneRedundancy = cache.getRegion(PR_ONE_REDUNDANCY); |
| for (int k = 0; k < 10; k++) { |
| regionOneRedundancy.put(k, k); |
| } |
| }); |
| |
| vm0.invoke(() -> validateBucketScope(PR_ZERO_REDUNDANCY, PR_ONE_REDUNDANCY)); |
| vm1.invoke(() -> validateBucketScope(PR_ZERO_REDUNDANCY, PR_ONE_REDUNDANCY)); |
| } |
| |
| private void validateBucketScope(String prZeroRedundancy, String prSingleRedundancy) { |
| Cache cache = getCache(); |
| |
| PartitionedRegion regionZeroRedundancy = (PartitionedRegion) cache.getRegion(prZeroRedundancy); |
| |
| for (BucketRegion bucket : regionZeroRedundancy.getDataStore().getLocalBucket2RegionMap() |
| .values()) { |
| assertThat(bucket.getAttributes().getScope().isDistributedAck()).isTrue(); |
| } |
| |
| PartitionedRegion regionOneRedundancy = (PartitionedRegion) cache.getRegion(prSingleRedundancy); |
| |
| for (Region bucket : regionOneRedundancy.getDataStore().getLocalBucket2RegionMap().values()) { |
| assertThat(bucket.getAttributes().getDataPolicy()).isSameAs(DataPolicy.REPLICATE); |
| } |
| } |
| |
| private void createPRs(String prZeroRedundancy, String prOneRedundancy) { |
| Cache cache = getCache(); |
| |
| // RedundantCopies = 0 , Scope = DISTRIBUTED_ACK |
| createPartitionedRegion(cache, prZeroRedundancy, 200, 0, totalNumBuckets); |
| // RedundantCopies > 0 , Scope = DISTRIBUTED_ACK |
| createPartitionedRegion(cache, prOneRedundancy, 200, 1, totalNumBuckets); |
| } |
| |
| private Object validateBucketsOnNode(int whichRegion) { |
| int containsNode = 0; |
| Cache cache = getCache(); |
| PartitionedRegion partitionedRegion = |
| (PartitionedRegion) cache.getRegion(prNamePrefix + whichRegion); |
| RegionAdvisor regionAdvisor = partitionedRegion.getRegionAdvisor(); |
| |
| try { |
| for (int bucketId : regionAdvisor.getBucketSet()) { |
| Set<InternalDistributedMember> nodeList = regionAdvisor.getBucketOwners(bucketId); |
| if (nodeList != null && nodeList.contains(partitionedRegion.getMyId())) { |
| containsNode++; |
| } |
| } |
| } catch (NoSuchElementException ignored) { |
| } |
| |
| return containsNode; |
| } |
| |
| private Object validateLocalBucket2RegionMapSize(int whichRegion) { |
| int size = 0; |
| Cache cache = getCache(); |
| PartitionedRegion partitionedRegion = |
| (PartitionedRegion) cache.getRegion(prNamePrefix + whichRegion); |
| PartitionedRegionDataStore dataStore = partitionedRegion.getDataStore(); |
| if (dataStore != null) { |
| size = dataStore.getBucketsManaged(); |
| } |
| return size; |
| } |
| |
| private void putsInDatastoreVM0() { |
| Cache cache = getCache(); |
| for (int i = 0; i < numberOfRegions; i++) { |
| Region<String, String> region = cache.getRegion(prNamePrefix + i); |
| for (int k = 0; k < 10; k++) { |
| region.put(i + prNamePrefix + k, prNamePrefix + k); |
| } |
| } |
| } |
| |
| private void createPRsAndAwaitRecovery(int localMaxMemory, int redundancy, int totalNumBuckets) |
| throws InterruptedException { |
| CountDownLatch recoveryDone = new CountDownLatch(numberOfRegions); |
| |
| ResourceObserver waitForRecovery = new ResourceObserverAdapter() { |
| @Override |
| public void rebalancingOrRecoveryFinished(Region region) { |
| recoveryDone.countDown(); |
| } |
| }; |
| |
| InternalResourceManager.setResourceObserver(waitForRecovery); |
| String originalValue = setSystemProperty(RETRY_TIMEOUT_PROPERTY, RETRY_TIMEOUT_VALUE); |
| try { |
| Cache cache = getCache(); |
| for (int i = 0; i < numberOfRegions; i++) { |
| String regionName = prNamePrefix + i; |
| createPartitionedRegion(cache, regionName, localMaxMemory, redundancy, totalNumBuckets); |
| } |
| assertThat(recoveryDone.await(1, MINUTES)).isTrue(); |
| } finally { |
| InternalResourceManager.setResourceObserver(null); |
| setSystemProperty(RETRY_TIMEOUT_PROPERTY, originalValue); |
| } |
| } |
| |
| private String setSystemProperty(String property, String value) { |
| if (value == null) { |
| return System.clearProperty(property); |
| } else { |
| return System.setProperty(property, value); |
| } |
| } |
| |
| private void validateEntries(boolean value) { |
| Cache cache = getCache(); |
| PartitionedRegion region = (PartitionedRegion) cache.getRegion(regionName); |
| for (int i = region.getTotalNumberOfBuckets() * 2; i >= 0; --i) { |
| assertThat(region).containsKey(i); |
| assertThat(region.get(i)).isEqualTo(value); |
| } |
| } |
| |
| private void createPartitionedRegion() throws InterruptedException { |
| CountDownLatch rebalancingFinished = new CountDownLatch(1); |
| |
| ResourceObserver waitForRebalancing = new ResourceObserverAdapter() { |
| @Override |
| public void rebalancingOrRecoveryFinished(Region region) { |
| rebalancingFinished.countDown(); |
| } |
| }; |
| |
| InternalResourceManager.setResourceObserver(waitForRebalancing); |
| try { |
| Cache cache = getCache(); |
| Region partitionedRegion = createPartitionedRegion(cache, regionName, 20, 1, totalNumBuckets); |
| assertThat(rebalancingFinished.await(1, MINUTES)).isTrue(); |
| assertThat(partitionedRegion).isNotNull(); |
| } finally { |
| InternalResourceManager.setResourceObserver(null); |
| } |
| } |
| |
| private PartitionedRegion createPartitionedRegion(Cache cache, String regionName, |
| int localMaxMemory, int redundancy, int totalNumBuckets) { |
| PartitionAttributesFactory partitionAttributesFactory = new PartitionAttributesFactory(); |
| partitionAttributesFactory.setLocalMaxMemory(localMaxMemory); |
| partitionAttributesFactory.setRedundantCopies(redundancy); |
| partitionAttributesFactory.setTotalNumBuckets(totalNumBuckets); |
| |
| RegionFactory regionFactory = cache.createRegionFactory(RegionShortcut.PARTITION); |
| regionFactory.setPartitionAttributes(partitionAttributesFactory.create()); |
| return (PartitionedRegion) regionFactory.create(regionName); |
| } |
| } |