[
https://issues.apache.org/jira/browse/CASSANDRA-12368?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15405062#comment-15405062
]
Paulo Motta edited comment on CASSANDRA-12368 at 8/3/16 12:24 AM:
------------------------------------------------------------------
added
[dtest|https://github.com/pauloricardomg/cassandra-dtest/commit/9cba01af42caa7578bf4c7f0e31ad032579be9a8]
to reproduce issue.
basic idea is to modify original
{{read_repair_test.py:TestReadRepair.alter_rf_and_run_read_repair_test}}, which
queries at {{CL=ALL}} to trigger read repair, to instead set
{{read_repair_chance}} to 1.0 and query at {{CL=ONE}}, verifying rows were
successfully repaired afterwards.
The test passes on 2.1 and 2.2 but fails on 3.0+
was (Author: pauloricardomg):
will post dtest soon.
> Background read repair not fixing replicas
> ------------------------------------------
>
> Key: CASSANDRA-12368
> URL: https://issues.apache.org/jira/browse/CASSANDRA-12368
> Project: Cassandra
> Issue Type: Bug
> Components: Coordination, Local Write-Read Paths
> Reporter: Paulo Motta
> Fix For: 3.x
>
>
> On 3.x background read repair is detecting mismatch and performing background
> reads, but for some reason it's not actually repairing replicas. The problem
> apparently does not occur on 2.2-.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
