[
https://issues.apache.org/jira/browse/CASSANDRA-2582?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13063368#comment-13063368
]
Timo Nentwig commented on CASSANDRA-2582:
-----------------------------------------
I certainly agree that this is this is a bad hack but it's IMHO still better
than running thru potentially very large JSON files (which likely cannot be
loaded into RAM) just to strip out some stuff.
Actually the cleanest way was if it was possible to set up the archive
node/cluster as another DC that the live cassandra nodes also replicate to and
the archive node/cluster to have a overriding local TTL policy (i.e. no TTL).
> sstable2json: make TTL optional, possibility of incremental dumps
> -----------------------------------------------------------------
>
> Key: CASSANDRA-2582
> URL: https://issues.apache.org/jira/browse/CASSANDRA-2582
> Project: Cassandra
> Issue Type: Improvement
> Components: Tools
> Reporter: Timo Nentwig
> Priority: Trivial
> Fix For: 0.8.2
>
> Attachments: nottl-and-lastmodified-0.7.5.patch,
> nottl-and-lastmodified.patch
>
>
> Data on live cluster is supposed to expire automatically but exported to an
> archive cluster beforehand. In order to prevent expiration on the archive
> cluster add an option to sstable2json to ignore TTL data. Furthermore add an
> lastModified option that ignores older columns for incremental archiving.
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
