[
https://issues.apache.org/jira/browse/FLINK-15368?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17009817#comment-17009817
]
Stephan Ewen commented on FLINK-15368:
--------------------------------------
Concerning the performance implications of the WriteBufferManager, can you
explain that a bit more? When is memory reserved? Is that every 1MB when a
"dummy block" is inserted into the cache to represent the used write buffer
memory?
Concerning the write buffer ration: Do you have any insights from running this
during tests which value is on average more beneficial for performance?
Concerning fewer column families: I recall that in the previous discussion we
left off with a question of whether this might depend on the layout/clustering
(order of components in the key). Maybe I misunderstood this, but it sounded
like it was not fully concluded, yet.
> Add end-to-end test for controlling RocksDB memory usage
> --------------------------------------------------------
>
> Key: FLINK-15368
> URL: https://issues.apache.org/jira/browse/FLINK-15368
> Project: Flink
> Issue Type: Sub-task
> Components: Runtime / State Backends
> Affects Versions: 1.10.0
> Reporter: Yu Li
> Assignee: Yun Tang
> Priority: Critical
> Fix For: 1.10.0
>
>
> We need to add an end-to-end test to make sure the RocksDB memory usage
> control works well, especially under the slot sharing case.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
