Barrett Oglesby created GEODE-9030:
--------------------------------------
Summary: The PartitionedIndex arbitraryBucketIndex doesn't get
reset when the BucketRegion defining it is moved
Key: GEODE-9030
URL: https://issues.apache.org/jira/browse/GEODE-9030
Project: Geode
Issue Type: Bug
Components: querying
Reporter: Barrett Oglesby
This causes a RegionDestroyedException like this when executing a query
containing a != clause:
{noformat}
Exception in thread "main"
org.apache.geode.cache.client.ServerOperationException: remote server on
10.166.145.16(client:27461:loner):58776:dfd3ba27:client: While performing a
remote query
at
org.apache.geode.cache.client.internal.AbstractOp.processChunkedResponse(AbstractOp.java:342)
at
org.apache.geode.cache.client.internal.QueryOp$QueryOpImpl.processResponse(QueryOp.java:168)
at
org.apache.geode.cache.client.internal.AbstractOp.processResponse(AbstractOp.java:224)
at
org.apache.geode.cache.client.internal.AbstractOp.attemptReadResponse(AbstractOp.java:197)
at
org.apache.geode.cache.client.internal.AbstractOp.attempt(AbstractOp.java:384)
at
org.apache.geode.cache.client.internal.ConnectionImpl.execute(ConnectionImpl.java:284)
at
org.apache.geode.cache.client.internal.pooling.PooledConnection.execute(PooledConnection.java:355)
at
org.apache.geode.cache.client.internal.OpExecutorImpl.executeWithPossibleReAuthentication(OpExecutorImpl.java:756)
at
org.apache.geode.cache.client.internal.OpExecutorImpl.execute(OpExecutorImpl.java:142)
at
org.apache.geode.cache.client.internal.OpExecutorImpl.execute(OpExecutorImpl.java:112)
at
org.apache.geode.cache.client.internal.PoolImpl.execute(PoolImpl.java:797)
at
org.apache.geode.cache.client.internal.QueryOp.execute(QueryOp.java:59)
at
org.apache.geode.cache.client.internal.ServerProxy.query(ServerProxy.java:59)
at
org.apache.geode.cache.query.internal.DefaultQuery.executeOnServer(DefaultQuery.java:327)
at
org.apache.geode.cache.query.internal.DefaultQuery.execute(DefaultQuery.java:215)
at
org.apache.geode.cache.query.internal.DefaultQuery.execute(DefaultQuery.java:197)
Caused by: org.apache.geode.cache.query.QueryInvocationTargetException: The
Region on which query is executed may have been
destroyed.BucketRegion[path='/__PR/_B__trade_0;serial=12;primary=false]
at
org.apache.geode.internal.cache.PRQueryProcessor.executeQueryOnBuckets(PRQueryProcessor.java:264)
at
org.apache.geode.internal.cache.PRQueryProcessor.executeSequentially(PRQueryProcessor.java:214)
at
org.apache.geode.internal.cache.PRQueryProcessor.executeQuery(PRQueryProcessor.java:124)
at
org.apache.geode.internal.cache.partitioned.QueryMessage.operateOnPartitionedRegion(QueryMessage.java:210)
Caused by: org.apache.geode.cache.RegionDestroyedException:
BucketRegion[path='/__PR/_B__trade_0;serial=12;primary=false]
at
org.apache.geode.internal.cache.LocalRegion.checkRegionDestroyed(LocalRegion.java:7352)
at
org.apache.geode.internal.cache.LocalRegion.checkReadiness(LocalRegion.java:2757)
at
org.apache.geode.internal.cache.BucketRegion.checkReadiness(BucketRegion.java:1437)
at
org.apache.geode.internal.cache.LocalRegion.size(LocalRegion.java:8313)
at
org.apache.geode.cache.query.internal.index.CompactRangeIndex.getSizeEstimate(CompactRangeIndex.java:331)
at
org.apache.geode.cache.query.internal.CompiledComparison.getSizeEstimate(CompiledComparison.java:337)
at
org.apache.geode.cache.query.internal.GroupJunction.organizeOperands(GroupJunction.java:146)
at
org.apache.geode.cache.query.internal.AbstractGroupOrRangeJunction.filterEvaluate(AbstractGroupOrRangeJunction.java:148)
at
org.apache.geode.cache.query.internal.CompiledJunction.filterEvaluate(CompiledJunction.java:190)
at
org.apache.geode.cache.query.internal.CompiledSelect.evaluate(CompiledSelect.java:538)
at
org.apache.geode.cache.query.internal.CompiledSelect.evaluate(CompiledSelect.java:53)
at
org.apache.geode.cache.query.internal.DefaultQuery.executeUsingContext(DefaultQuery.java:357)
at
org.apache.geode.internal.cache.PRQueryProcessor.executeQueryOnBuckets(PRQueryProcessor.java:248)
{noformat}
Here is an example query that fails:
{noformat}
SELECT * FROM /trade WHERE arrangementId = 'aId_1' AND tradeStatus.toString()
!= 'CLOSED'
{noformat}
Here is a test that reproduces it:
* start one server with region configured as PARTITION with:
** 2 buckets
** PartitionResolver that puts the first entry in bucket 0, every other entry
in bucket 1
* load N entries
* the index in bucket 0 becomes the arbitraryBucketIndex
* start a second server
* rebalance
* bucket 0 moves from the first server to the second server
* run the query
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
