You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@cassandra.apache.org by "T Jake Luciani (JIRA)" <ji...@apache.org> on 2014/06/05 19:39:03 UTC
[jira] [Commented] (CASSANDRA-2527) Add ability to snapshot data as
input to hadoop jobs
[ https://issues.apache.org/jira/browse/CASSANDRA-2527?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14019018#comment-14019018 ]
T Jake Luciani commented on CASSANDRA-2527:
-------------------------------------------
This might be a bit off the wall but my idea was to make a mock read only hdfs compatible file system.
The structure of which would look like:
{code}
/nodes
/ ip1
/snapshot1/
/keyspace1/
SStable1-data.db
.....
/snapshot2/
...
/ ip2
...
{code}
Since this data is immutable there would be no namenode service, it would simply query the nodes for the file info.
If the mapper runs on a non-local node we use a sendfile() copy mechanism.
> Add ability to snapshot data as input to hadoop jobs
> ----------------------------------------------------
>
> Key: CASSANDRA-2527
> URL: https://issues.apache.org/jira/browse/CASSANDRA-2527
> Project: Cassandra
> Issue Type: New Feature
> Reporter: Jeremy Hanna
> Assignee: Joshua McKenzie
> Priority: Minor
> Labels: hadoop
> Fix For: 3.0
>
>
> It is desirable to have immutable inputs to hadoop jobs for the duration of the job. That way re-execution of individual tasks do not alter the output. One way to accomplish this would be to snapshot the data that is used as input to a job.
--
This message was sent by Atlassian JIRA
(v6.2#6252)