[
https://issues.apache.org/jira/browse/CASSANDRA-20752?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Michael Semb Wever updated CASSANDRA-20752:
-------------------------------------------
Fix Version/s: 5.1
(was: 6.x)
> Flush SAI segment builder when current sstable writer is switched
> -----------------------------------------------------------------
>
> Key: CASSANDRA-20752
> URL: https://issues.apache.org/jira/browse/CASSANDRA-20752
> Project: Apache Cassandra
> Issue Type: Improvement
> Components: Feature/SAI
> Reporter: Zhao Yang
> Assignee: Pranav
> Priority: Normal
> Fix For: 5.0.6, 5.1
>
> Attachments: ci_summary.html
>
> Time Spent: 3h 40m
> Remaining Estimate: 0h
>
> With sharded compaction, sstable writer may be switched when reaching
> boundary in `SSTableRewriter#switchWriter`. But the switched writer's segment
> builders are not flushed until all sharded are written and committed.
> This unccessarily retains memory causing smaller segments and may even cause
> OOM due to segment builder's min-flush-size.
> We can extend `SSTableFlushObserver` to be notified on sstable switched
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
