[
https://issues.apache.org/jira/browse/HDFS-8070?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Colin Patrick McCabe reassigned HDFS-8070:
------------------------------------------
Assignee: Colin Patrick McCabe (was: Kihwal Lee)
> Pre-HDFS-7915 DFSClient cannot use short circuit on post-HDFS-7915 DataNode
> ---------------------------------------------------------------------------
>
> Key: HDFS-8070
> URL: https://issues.apache.org/jira/browse/HDFS-8070
> Project: Hadoop HDFS
> Issue Type: Bug
> Components: caching
> Affects Versions: 2.8.0
> Reporter: Gopal V
> Assignee: Colin Patrick McCabe
>
> HDFS ShortCircuitShm layer keeps the task locked up during multi-threaded
> split-generation.
> I hit this immediately after I upgraded the data, so I wonder if the
> ShortCircuitShim wire protocol has trouble when 2.8.0 DN talks to a 2.7.0
> Client?
> {code}
> 2015-04-06 00:04:30,780 INFO [ORC_GET_SPLITS #3] orc.OrcInputFormat: ORC
> pushdown predicate: leaf-0 = (IS_NULL ss_sold_date_sk)
> expr = (not leaf-0)
> 2015-04-06 00:04:30,781 ERROR [ShortCircuitCache_SlotReleaser]
> shortcircuit.ShortCircuitCache: ShortCircuitCache(0x29e82045): failed to
> release short-circuit shared memory slot Slot(slotIdx=2,
> shm=DfsClientShm(a86ee34576d93c4964005d90b0d97c38)) by sending
> ReleaseShortCircuitAccessRequestProto to /grid/0/cluster/hdfs/dn_socket.
> Closing shared memory segment.
> java.io.IOException: ERROR_INVALID: there is no shared memory segment
> registered with shmId a86ee34576d93c4964005d90b0d97c38
> at
> org.apache.hadoop.hdfs.shortcircuit.ShortCircuitCache$SlotReleaser.run(ShortCircuitCache.java:208)
> at
> java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:511)
> at java.util.concurrent.FutureTask.run(FutureTask.java:266)
> at
> java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$201(ScheduledThreadPoolExecutor.java:180)
> at
> java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:293)
> at
> java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1142)
> at
> java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:617)
> at java.lang.Thread.run(Thread.java:745)
> 2015-04-06 00:04:30,781 INFO [ORC_GET_SPLITS #5] orc.OrcInputFormat: ORC
> pushdown predicate: leaf-0 = (IS_NULL ss_sold_date_sk)
> expr = (not leaf-0)
> 2015-04-06 00:04:30,781 WARN [ShortCircuitCache_SlotReleaser]
> shortcircuit.DfsClientShmManager: EndpointShmManager(172.19.128.60:50010,
> parent=ShortCircuitShmManager(5e763476)): error shutting down shm: got
> IOException calling shutdown(SHUT_RDWR)
> java.nio.channels.ClosedChannelException
> at
> org.apache.hadoop.util.CloseableReferenceCount.reference(CloseableReferenceCount.java:57)
> at
> org.apache.hadoop.net.unix.DomainSocket.shutdown(DomainSocket.java:387)
> at
> org.apache.hadoop.hdfs.shortcircuit.DfsClientShmManager$EndpointShmManager.shutdown(DfsClientShmManager.java:378)
> at
> org.apache.hadoop.hdfs.shortcircuit.ShortCircuitCache$SlotReleaser.run(ShortCircuitCache.java:223)
> at
> java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:511)
> at java.util.concurrent.FutureTask.run(FutureTask.java:266)
> at
> java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$201(ScheduledThreadPoolExecutor.java:180)
> at
> java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:293)
> at
> java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1142)
> at
> java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:617)
> at java.lang.Thread.run(Thread.java:745)
> 2015-04-06 00:04:30,783 INFO [ORC_GET_SPLITS #7] orc.OrcInputFormat: ORC
> pushdown predicate: leaf-0 = (IS_NULL cs_sold_date_sk)
> expr = (not leaf-0)
> 2015-04-06 00:04:30,785 ERROR [ShortCircuitCache_SlotReleaser]
> shortcircuit.ShortCircuitCache: ShortCircuitCache(0x29e82045): failed to
> release short-circuit shared memory slot Slot(slotIdx=4,
> shm=DfsClientShm(a86ee34576d93c4964005d90b0d97c38)) by sending
> ReleaseShortCircuitAccessRequestProto to /grid/0/cluster/hdfs/dn_socket.
> Closing shared memory segment.
> java.io.IOException: ERROR_INVALID: there is no shared memory segment
> registered with shmId a86ee34576d93c4964005d90b0d97c38
> at
> org.apache.hadoop.hdfs.shortcircuit.ShortCircuitCache$SlotReleaser.run(ShortCircuitCache.java:208)
> at
> java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:511)
> at java.util.concurrent.FutureTask.run(FutureTask.java:266)
> at
> java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$201(ScheduledThreadPoolExecutor.java:180)
> at
> java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:293)
> at
> java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1142)
> at
> java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:617)
> at java.lang.Thread.run(Thread.java:745)
> {code}
> Looks like a double free-fd condition?
> {code}
> 2015-04-02 18:58:47,653 [DataXceiver for client
> unix:/grid/0/cluster/hdfs/dn_socket [Passing file descriptors for block
> BP-942051088-172.18.0.41-1370508013893:blk_1076973408_1099515627985]] INFO
> org.apache.hadoop.hdfs.server.datanode.DataNode: Unregistering
> SlotId(3bd7fd9aed791e95acfb5034e6617d83:0) because the
> requestShortCircuitFdsForRead operation failed.
> 2015-04-02 18:58:47,653 [DataXceiver for client
> unix:/grid/0/cluster/hdfs/dn_socket [Passing file descriptors for block
> BP-942051088-<ip>-1370508013893:blk_1076973408_1099515627985]] INFO
> org.apache.hadoop.hdfs.server.datanode.DataNode.clienttrace: src: 127.0.0.1,
> dest: 127.0.0.1, op: REQUEST_SHORT_CIRCUIT_FDS, blockid: 1076973408, srvID:
> ba7b6f19-47e0-4b86-af50-23981649318c, success: false
> 2015-04-02 18:58:47,654 [DataXceiver for client
> unix:/grid/0/cluster/hdfs/dn_socket [Passing file descriptors for block
> BP-942051088-172.18.0.41-1370508013893:blk_1076973408_1099515627985]] ERROR
> org.apache.hadoop.hdfs.server.datanode.DataNode:
> cn060-10.l42scl.hortonworks.com:50010:DataXceiver error processing
> REQUEST_SHORT_CIRCUIT_FDS operation src: unix:/grid/0/cluster/hdfs/dn_socket
> dst: <local>
> java.io.EOFException
> at
> org.apache.hadoop.hdfs.server.datanode.DataXceiver.requestShortCircuitFds(DataXceiver.java:352)
> at
> org.apache.hadoop.hdfs.protocol.datatransfer.Receiver.opRequestShortCircuitFds(Receiver.java:187)
> at
> org.apache.hadoop.hdfs.protocol.datatransfer.Receiver.processOp(Receiver.java:89)
> at
> org.apache.hadoop.hdfs.server.datanode.DataXceiver.run(DataXceiver.java:251)
> at java.lang.Thread.run(Thread.java:745)
> {code}
> Investigating more, since the exact exception from the DataNode call is not
> logged.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
