[
https://issues.apache.org/jira/browse/CASSANDRA-10357?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14903410#comment-14903410
]
T Jake Luciani commented on CASSANDRA-10357:
--------------------------------------------
I asked [~Stefania] to verify the logic in forceRepair and she agrees with the
code. If you could add some comments on that code it would be helpful for
unfamiliar readers.
The downside of this is we will run the check code on every sstable for all
time if the boundaries are ok until it's compacted. I'd rather have a way to
only check this once then follow on versions would never check.
Nit: The magic constant you added, could you make that a const and refer to it
vs using the actual number?
> mmap file boundary selection is broken for some large files
> ------------------------------------------------------------
>
> Key: CASSANDRA-10357
> URL: https://issues.apache.org/jira/browse/CASSANDRA-10357
> Project: Cassandra
> Issue Type: Bug
> Components: Core
> Reporter: Benedict
> Assignee: Benedict
> Fix For: 2.1.10
>
>
> If an early open interval occurs to close to an mmap boundary, the boundary
> can be lost. Patch available
> [here|https://github.com/belliottsmith/cassandra/tree/mmap-boundaries].
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
