[
https://issues.apache.org/jira/browse/FLINK-19936?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Guowei Ma updated FLINK-19936:
------------------------------
Description:
In the streaming execution mode there are two uncertain things:
# The order of receiving the checkpoint complete.(`Committer` receive first or
`GlobalCommitter` receive first)
# Whether the operator can accept the last checkpoint
This patch does following two things to resolve `SinkITCase`'s unstable owning
to these two uncertainty
# Changing how to verifying the output of `GlobalCommitter` for the case 1
# Let the `FiniteTestSource` exit only when the Committer/GlobalCommitter
received all the elements.
was:
In the streaming execution mode there are two uncertain things:
# The order of receiving the checkpoint complete.(`Committer` receive first or
`GlobalCommitter` receive first)
# Whether the operator can accept the last checkpoint
These lead to the `SinkITCase` unstable.
This pr makes the `SinkITCase`s unstable by “resolving” these two uncertain
stuff.
> Make SinkITCase stable
> ----------------------
>
> Key: FLINK-19936
> URL: https://issues.apache.org/jira/browse/FLINK-19936
> Project: Flink
> Issue Type: Sub-task
> Components: API / DataStream
> Reporter: Guowei Ma
> Priority: Major
>
> In the streaming execution mode there are two uncertain things:
> # The order of receiving the checkpoint complete.(`Committer` receive first
> or `GlobalCommitter` receive first)
> # Whether the operator can accept the last checkpoint
> This patch does following two things to resolve `SinkITCase`'s unstable
> owning to these two uncertainty
> # Changing how to verifying the output of `GlobalCommitter` for the case 1
> # Let the `FiniteTestSource` exit only when the Committer/GlobalCommitter
> received all the elements.
>
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
