[infinispan-issues] [JBoss JIRA] Commented: (ISPN-1091) HotRod clients should failover if encountering a SuspectException

[Galder Zamarreño (JIRA)]

    [ https://issues.jboss.org/browse/ISPN-1091?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12601237#comment-12601237 ] 

Galder Zamarreño commented on ISPN-1091:
----------------------------------------

To sump up some of the discussion happening in JBPAPP-6473, the server should not be the one doing the failover if it receives a SuspectException or similar out of a cache invocation. It should launder it and bubble it up to the client who's the one ultimately making the decision to failover or not. 

Also, as rightly pointed out by Michal, we should check what happens with repl timeout exception. AFAIK, this should have been covered under the 0x86 command timed out status code in http://community.jboss.org/docs/DOC-14421 and if that's the case, it should failover already. If it's not, it needs changing.

> HotRod clients should failover if encountering a SuspectException
> -----------------------------------------------------------------
>
>                 Key: ISPN-1091
>                 URL: https://issues.jboss.org/browse/ISPN-1091
>             Project: Infinispan
>          Issue Type: Bug
>          Components: Cache Server
>    Affects Versions: 4.2.1.FINAL
>            Reporter: Galder Zamarreño
>            Assignee: Galder Zamarreño
>             Fix For: 4.2.2.BETA1, 4.2.2.FINAL, 5.0.0.CR2
>
>
> As Michal rightly indicated in JBPAPP-6473, Hot Rod clients need to be able to deal with SuspectException errors and failover rather than bubbling up. In the future the server will make sure the right error code is returned, but in the absence of that, the error message can be inspected.

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