[
https://issues.apache.org/jira/browse/HBASE-9863?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13812539#comment-13812539
]
Jimmy Xiang commented on HBASE-9863:
------------------------------------
I think start() should be synchronized too. It's better to make sure it won't
be called more than once also. For 'NamespaceDescriptor get(String name)',
yes, the semantics is the same. After the change, we don't check if
isTableAvailableAndInitialized. So zkNamespaceManager could be uninitialized.
> Intermittently
> TestZooKeeper#testRegionAssignmentAfterMasterRecoveryDueToZKExpiry hangs
> ---------------------------------------------------------------------------------------
>
> Key: HBASE-9863
> URL: https://issues.apache.org/jira/browse/HBASE-9863
> Project: HBase
> Issue Type: Bug
> Reporter: Ted Yu
> Assignee: Ted Yu
> Attachments: 9863-v1.txt, 9863-v2.txt, 9863-v3.txt
>
>
> TestZooKeeper#testRegionAssignmentAfterMasterRecoveryDueToZKExpiry sometimes
> hung.
> Here were two recent occurrences:
> https://builds.apache.org/job/PreCommit-HBASE-Build/7676/console
> https://builds.apache.org/job/PreCommit-HBASE-Build/7671/console
> There were 9 occurrences of the following in both stack traces:
> {code}
> "FifoRpcScheduler.handler1-thread-5" daemon prio=10 tid=0x09df8800 nid=0xc17
> waiting for monitor entry [0x6fdf8000]
> java.lang.Thread.State: BLOCKED (on object monitor)
> at
> org.apache.hadoop.hbase.master.TableNamespaceManager.isTableAvailableAndInitialized(TableNamespaceManager.java:250)
> - waiting to lock <0x7f69b5f0> (a
> org.apache.hadoop.hbase.master.TableNamespaceManager)
> at
> org.apache.hadoop.hbase.master.HMaster.isTableNamespaceManagerReady(HMaster.java:3146)
> at
> org.apache.hadoop.hbase.master.HMaster.getNamespaceDescriptor(HMaster.java:3105)
> at org.apache.hadoop.hbase.master.HMaster.createTable(HMaster.java:1743)
> at org.apache.hadoop.hbase.master.HMaster.createTable(HMaster.java:1782)
> at
> org.apache.hadoop.hbase.protobuf.generated.MasterProtos$MasterService$2.callBlockingMethod(MasterProtos.java:38221)
> at org.apache.hadoop.hbase.ipc.RpcServer.call(RpcServer.java:1983)
> at org.apache.hadoop.hbase.ipc.CallRunner.run(CallRunner.java:92)
> {code}
> The test hung here:
> {code}
> "pool-1-thread-1" prio=10 tid=0x74f7b800 nid=0x5aa5 in Object.wait()
> [0x74efe000]
> java.lang.Thread.State: TIMED_WAITING (on object monitor)
> at java.lang.Object.wait(Native Method)
> - waiting on <0xcc848348> (a org.apache.hadoop.hbase.ipc.RpcClient$Call)
> at org.apache.hadoop.hbase.ipc.RpcClient.call(RpcClient.java:1436)
> - locked <0xcc848348> (a org.apache.hadoop.hbase.ipc.RpcClient$Call)
> at
> org.apache.hadoop.hbase.ipc.RpcClient.callBlockingMethod(RpcClient.java:1654)
> at
> org.apache.hadoop.hbase.ipc.RpcClient$BlockingRpcChannelImplementation.callBlockingMethod(RpcClient.java:1712)
> at
> org.apache.hadoop.hbase.protobuf.generated.MasterProtos$MasterService$BlockingStub.createTable(MasterProtos.java:40372)
> at
> org.apache.hadoop.hbase.client.HConnectionManager$HConnectionImplementation$5.createTable(HConnectionManager.java:1931)
> at org.apache.hadoop.hbase.client.HBaseAdmin$2.call(HBaseAdmin.java:598)
> at org.apache.hadoop.hbase.client.HBaseAdmin$2.call(HBaseAdmin.java:594)
> at
> org.apache.hadoop.hbase.client.RpcRetryingCaller.callWithRetries(RpcRetryingCaller.java:116)
> - locked <0x7faa26d0> (a org.apache.hadoop.hbase.client.RpcRetryingCaller)
> at
> org.apache.hadoop.hbase.client.RpcRetryingCaller.callWithRetries(RpcRetryingCaller.java:94)
> - locked <0x7faa26d0> (a org.apache.hadoop.hbase.client.RpcRetryingCaller)
> at
> org.apache.hadoop.hbase.client.HBaseAdmin.executeCallable(HBaseAdmin.java:3124)
> at
> org.apache.hadoop.hbase.client.HBaseAdmin.createTableAsync(HBaseAdmin.java:594)
> at
> org.apache.hadoop.hbase.client.HBaseAdmin.createTable(HBaseAdmin.java:485)
> at
> org.apache.hadoop.hbase.TestZooKeeper.testRegionAssignmentAfterMasterRecoveryDueToZKExpiry(TestZooKeeper.java:486)
> {code}
--
This message was sent by Atlassian JIRA
(v6.1#6144)
