[
https://issues.apache.org/jira/browse/CASSANDRA-13703?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Branimir Lambov updated CASSANDRA-13703:
----------------------------------------
Reviewer: Alex Petrov
Status: Patch Available (was: In Progress)
Patch [here|https://github.com/blambov/cassandra/tree/13703].
Changes the max size threshold to be non-inclusive so that
{{min_compress_ratio: 1}} can work (this would be a breaking change, but
CASSANRDA-10520 is not part of any release yet), and adds an error for values
between 0 and 1 exclusive.
Unit tests are clean, dtests have two failures that appear to be flaky on trunk:
{{repair_tests.incremental_repair_test.TestIncRepair.multiple_repair_test}}
{{bootstrap_test.TestBootstrap.consistent_range_movement_false_with_rf1_should_succeed_test}}
> Using min_compress_ratio <= 1 causes corruption
> -----------------------------------------------
>
> Key: CASSANDRA-13703
> URL: https://issues.apache.org/jira/browse/CASSANDRA-13703
> Project: Cassandra
> Issue Type: Bug
> Reporter: Branimir Lambov
> Assignee: Branimir Lambov
> Attachments: patch
>
>
> This is because chunks written uncompressed end up below the compressed size
> threshold. Demonstrated by applying the attached patch meant to improve the
> testing of the 10520 changes, and running
> {{CompressedSequentialWriterTest.testLZ4Writer}}.
> The default {{min_compress_ratio: 0}} is not affected as it never writes
> uncompressed.
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscr...@cassandra.apache.org
For additional commands, e-mail: commits-h...@cassandra.apache.org
