You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@flink.apache.org by "Timo Walther (JIRA)" <ji...@apache.org> on 2018/04/19 13:15:00 UTC

[jira] [Resolved] (FLINK-8980) End-to-end test: BucketingSink

     [ https://issues.apache.org/jira/browse/FLINK-8980?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Timo Walther resolved FLINK-8980.
---------------------------------
    Resolution: Fixed

Fixed in 1.6.0: 42828e0219e50ea0cd4481346f24250a674fe596
Fixed in 1.5.0: 3df5c6685091cea6c50758769c37ad3d6390f94a

> End-to-end test: BucketingSink
> ------------------------------
>
>                 Key: FLINK-8980
>                 URL: https://issues.apache.org/jira/browse/FLINK-8980
>             Project: Flink
>          Issue Type: Sub-task
>          Components: Tests
>    Affects Versions: 1.5.0
>            Reporter: Till Rohrmann
>            Assignee: Timo Walther
>            Priority: Blocker
>             Fix For: 1.5.0
>
>
> In order to verify the {{BucketingSink}}, we should add an end-to-end test which verifies that the {{BucketingSink}} does not lose data under failures.
> An idea would be to have a CountUp job which simply counts up a counter which is persisted. The emitted values will be written to disk by the {{BucketingSink}}. Now we should kill randomly Flink processes (cluster entrypoint and TaskExecutors) to simulate failures. Even after these failures, the written files should contain the correct sequence of numbers.



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)