cassandra-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Thom Valley (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (CASSANDRA-11324) Make SSTABLESCRUB and related jobs multi-threaded / parallelize somehow
Date Thu, 02 Jun 2016 17:16:59 GMT

    [ https://issues.apache.org/jira/browse/CASSANDRA-11324?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15312686#comment-15312686
] 

Thom Valley commented on CASSANDRA-11324:
-----------------------------------------

I disagree.

SSTABLESCRUB is run on a keyspace or keyspace/table.

A single table can have hundreds of SSTABLEs, which _could_ be run in parallel.

Expecting a user to write a script to efficiently do basic maintenance is a serious usability
problem.

> Make SSTABLESCRUB and related jobs multi-threaded  / parallelize somehow
> ------------------------------------------------------------------------
>
>                 Key: CASSANDRA-11324
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-11324
>             Project: Cassandra
>          Issue Type: Improvement
>          Components: Tools
>            Reporter: Thom Valley
>            Priority: Minor
>
> SSTABLE scrub right now appears to be single threaded in offline mode (unsure about online).
 When scrubbing a node, especially a well provisioned one, the node sits mostly idle for potentially
hours while scrubbing.
> Making better use of available resources through parallelization / multi-threading of
the scrub process seems like a good idea.



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

Mime
View raw message