cassandra-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Jonathan Ellis (JIRA)" <j...@apache.org>
Subject [jira] Updated: (CASSANDRA-1641) auto-guessed memtable sizes are too high
Date Thu, 21 Oct 2010 05:28:16 GMT

     [ https://issues.apache.org/jira/browse/CASSANDRA-1641?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]

Jonathan Ellis updated CASSANDRA-1641:
--------------------------------------

    Attachment: 1641.txt

I'd like to introduce a dependency on number of user CFs, but that's not available until after
we've picked a value.  So this cuts the size by half, which feels scientific because on the
old default 1GB heap size this gives us memtable throughput of 64MB, which is also the old
default.

> auto-guessed memtable sizes are too high
> ----------------------------------------
>
>                 Key: CASSANDRA-1641
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-1641
>             Project: Cassandra
>          Issue Type: Bug
>          Components: Core
>    Affects Versions: 0.7 beta 2
>            Reporter: Jonathan Ellis
>            Assignee: Jonathan Ellis
>             Fix For: 0.7.0
>
>         Attachments: 1641.txt
>
>
> I've seen two cases now of the memtable sizes being too large, causing OOMing.  Too-small
memtables hurt performance, but too-large hurts worse when you start GC storming.

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


Mime
View raw message