[infinispan-issues] [JBoss JIRA] Commented: (ISPN-1273) hash distribution is not correctly transferred to hot rod client

[Michal Linhard (JIRA)]

    [ https://issues.jboss.org/browse/ISPN-1273?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12617224#comment-12617224 ] 

Michal Linhard commented on ISPN-1273:
--------------------------------------

this is actually a conceptual problem:
check out the code of readNewTopologyAndHash() in https://github.com/infinispan/infinispan/blob/5.0.0.CR8/client/hotrod-client/src/main/java/org/infinispan/client/hotrod/impl/operations/HotRodOperation.java

{code}
      for (int i = 0; i < clusterSize; i++) {
         String host = transport.readString();
         int port = transport.readUnsignedShort();
...
         int hashCode = transport.read4ByteInt();
         servers2HashCode.put(new InetSocketAddress(host, port), hashCode);
...
      }
{code}

the map should be hashCode2server, otherwise we rewrite host information with each new hash code

> hash distribution is not correctly transferred to hot rod client
> ----------------------------------------------------------------
>
>                 Key: ISPN-1273
>                 URL: https://issues.jboss.org/browse/ISPN-1273
>             Project: Infinispan
>          Issue Type: Bug
>    Affects Versions: 5.0.0.CR8
>            Reporter: Michal Linhard
>            Assignee: Galder Zamarreño
>
> I'm starting 4 ispn instances on interfaces test1-test4 with numVirtualNodes 512.
> when my client pings one of the servers for topology info - this is the contents of it's servers2HashCode map:
> 10213 = test1
> 10225 = test3
> 10235 = test2
> 10236 = test4
> i.e. it contains only one hashid per server.

--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira