[
https://issues.apache.org/jira/browse/CASSANDRA-9095?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14492429#comment-14492429
]
Branimir Lambov commented on CASSANDRA-9095:
--------------------------------------------
Not really, i.e. a fix for this will not help the test failure.
This bug is about the internal CL size tracking which influences when a flush
is initiated. AFAICS the test is failing because the actual file on disk has a
size that is different from expected. The real size is not that easy to predict
as it depends on the contents of the file. Using a fixed seed for the stress
call could help there, but I don't know if/how this can be done.
> Compressed commit log should measure compressed space used
> ----------------------------------------------------------
>
> Key: CASSANDRA-9095
> URL: https://issues.apache.org/jira/browse/CASSANDRA-9095
> Project: Cassandra
> Issue Type: Improvement
> Reporter: Branimir Lambov
> Assignee: Branimir Lambov
> Priority: Minor
>
> The commit log compression option introduced in CASSANDRA-6809 does not
> change the way space use is measure by the commitlog, meaning that it still
> measures the amount of space taken by the log segments before compression.
> It should measure the space taken on disk which in this case is the
> compressed amount.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
