You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@samza.apache.org by "Shanthoosh Venkataraman (JIRA)" <ji...@apache.org> on 2017/05/30 23:58:04 UTC

[jira] [Updated] (SAMZA-1308) Failure Testing. A processor dies while waiting for a barrier to complete.

     [ https://issues.apache.org/jira/browse/SAMZA-1308?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Shanthoosh Venkataraman updated SAMZA-1308:
-------------------------------------------
    Description: 
Tracks different failure scenarios testing around leader creating the barrier and other stream processors joining it. 

A) Death of a stream processor expected to join a zookeeper barrier should regenerate a new barrier with other alive stream processors. 
B) Zookeeper shutdown while waiting on the barrier should kill all stream applications. 
C) Should wait until all the expected stream processors joins the zookeeper barrier.
D) Should have fixed retries on failures when creating a barrier in zookeeper.
E) Barrier.expiry(when none of the stream application joins the barrier) should trigger the leader to create a new barrier with alive processors.

> Failure Testing. A processor dies while waiting for a barrier to complete.
> --------------------------------------------------------------------------
>
>                 Key: SAMZA-1308
>                 URL: https://issues.apache.org/jira/browse/SAMZA-1308
>             Project: Samza
>          Issue Type: Sub-task
>            Reporter: Boris Shkolnik
>            Assignee: Shanthoosh Venkataraman
>             Fix For: 0.13.0
>
>
> Tracks different failure scenarios testing around leader creating the barrier and other stream processors joining it. 
> A) Death of a stream processor expected to join a zookeeper barrier should regenerate a new barrier with other alive stream processors. 
> B) Zookeeper shutdown while waiting on the barrier should kill all stream applications. 
> C) Should wait until all the expected stream processors joins the zookeeper barrier.
> D) Should have fixed retries on failures when creating a barrier in zookeeper.
> E) Barrier.expiry(when none of the stream application joins the barrier) should trigger the leader to create a new barrier with alive processors.



--
This message was sent by Atlassian JIRA
(v6.3.15#6346)