[infinispan-issues] [JBoss JIRA] (ISPN-6086) Hadoop MapReduce task fails after the prefered ispn server stopped

[Gustavo Fernandes (JIRA)]

    [ https://issues.jboss.org/browse/ISPN-6086?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13157841#comment-13157841 ] 

Gustavo Fernandes edited comment on ISPN-6086 at 2/2/16 11:56 AM:
------------------------------------------------------------------

The issue is that the Splitter only accepts one host/ip, and if the server goes down after the job starts, but before the task actually contacts the Infinispan server, there's no way of discovering the other nodes. This can be solved by allowing propagating the list of servers to the tasks rather than only one. Created https://issues.jboss.org/browse/IHDP-2


was (Author: gustavonalle):
The issue is that the connector configuration only accepts one host/ip, and if the server goes down after the job starts, but before the task actually contacts the Infinispan server, there's no way of discovering the other nodes. This can be solved by allowing the configuration to have more than 1 server. Created https://issues.jboss.org/browse/IHDP-2

> Hadoop MapReduce task fails after the prefered ispn server stopped 
> -------------------------------------------------------------------
>
>                 Key: ISPN-6086
>                 URL: https://issues.jboss.org/browse/ISPN-6086
>             Project: Infinispan
>          Issue Type: Bug
>          Components: Hadoop Integration
>            Reporter: Anna Manukyan
>            Assignee: Gustavo Fernandes
>         Attachments: FailoverTest_when_prefered_stopped.log
>
>
> I have added tests to the infinispan-hadoop testsuite testing the load balancing of mapreduce task in case if the input server which is specified as a parameter to Hadoop configuration is stopped. 
> The test is performed on 4 node cluster. The data read is performed from Infinispan cache, the write is performed to HDFS. 
> There are 200000 entries in the cache. As soon as the Hadoop Job is submitted for performing map/reduce task, the preferred server is stopped. 
> As a result, the Hadoop Map/Reduce result map is empty.
> You can find the test execution log attached. 



--
This message was sent by Atlassian JIRA
(v6.4.11#64026)