cassandra-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Alexey Zotov (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (CASSANDRA-7493) Remove traces of rows_per_partition_to_cache
Date Tue, 02 Sep 2014 11:17:20 GMT

    [ https://issues.apache.org/jira/browse/CASSANDRA-7493?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14118101#comment-14118101
] 

Alexey Zotov commented on CASSANDRA-7493:
-----------------------------------------

Hmm.. why it was removed as a part of this cleanup? As for me _populate_io_cache_on_flush_
is irrelevant with caching mechanism because it uses internal OS capabilities for storing
data in-memory. 

Also I think it is quite strange to remove this feature since it can be very useful if dataset
fits memory. We used this option to have the whole dataset in memory and it gave us high performance
with low latency. 

What do you think about reverting the changes and starting a discussion regarding this option?


PS: 
I guess this option is not very popular because it does not have proper documentation. Maybe
it would be better to write an article (+ some info in official documentation) with detailed
explanation why and when it can be used than just remove it. If such approach works I can
help with it.

> Remove traces of rows_per_partition_to_cache
> --------------------------------------------
>
>                 Key: CASSANDRA-7493
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-7493
>             Project: Cassandra
>          Issue Type: Bug
>            Reporter: Marcus Eriksson
>            Assignee: Marcus Eriksson
>             Fix For: 2.1 rc3
>
>         Attachments: 0001-remove-rows_per_partition_to_cache.patch
>
>
> CASSANDRA-6745 changed the way we configure how many rows per partition we cache, some
old stuff was left in, clear it out



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Mime
View raw message