[
https://issues.apache.org/jira/browse/PHOENIX-4458?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16292119#comment-16292119
]
Ankit Singhal commented on PHOENIX-4458:
----------------------------------------
[~asko], As per your jstack, index writers are also using the same
pool(default) of handlers which is used for the scan.
So, until you upgrade to 4.12 (which has PHOENIX-3994 to automatically
configures the RPC controller factory for indexer), you need to manually set
the below configuration in hbase-site.xml of the server (ensure that these
configuration doesn't exist in hbase-site.xml of the client).
{code}
<property>
<name>hbase.region.server.rpc.scheduler.factory.class</name>
<value>org.apache.hadoop.hbase.ipc.PhoenixRpcSchedulerFactory</value>
<description>Factory to create the Phoenix RPC Scheduler that uses separate
queues for index and metadata updates</description>
</property>
<property>
<name>hbase.rpc.controllerfactory.class</name>
<value>org.apache.hadoop.hbase.ipc.controller.ServerRpcControllerFactory</value>
<description>Factory to create the Phoenix RPC Scheduler that uses separate
queues for index and metadata updates</description>
</property>
{code}
> Region dead lock when executing duplicate key upsert data table(have local
> index)
> ---------------------------------------------------------------------------------
>
> Key: PHOENIX-4458
> URL: https://issues.apache.org/jira/browse/PHOENIX-4458
> Project: Phoenix
> Issue Type: Bug
> Affects Versions: 4.11.0
> Reporter: asko
> Priority: Critical
> Attachments: RegionDeadLockTest.java, jstack
>
>
> The attach file *RegionDeadLockTest.java* can produce this bug after running
> a few minutes.
> The region will be hang that can not read/write.
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
