[
https://issues.apache.org/jira/browse/HBASE-2603?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12870452#action_12870452
]
Andrew Purtell commented on HBASE-2603:
---------------------------------------
Since this is real data, instead of generic Lucene and Katta (or hbasene +
Solr), consider anomaly detection:
* An online variant of
http://www2.berkeley.intel-research.net/~hling/research/minelog_sysml08.pdf
* SALSA/Mochi
(http://www.usenix.org/event/wasl08/tech/full_papers/tan/tan_html/),
incorporated into Apache Chukwa
(http://wiki.apache.org/hadoop/Anomaly_Detection_Framework_with_Chukwa)
Need to sanity check for gross failures but folding in the above could raise
red flags about possible subtle problems.
> full system application test scenario: event log
> ------------------------------------------------
>
> Key: HBASE-2603
> URL: https://issues.apache.org/jira/browse/HBASE-2603
> Project: Hadoop HBase
> Issue Type: Sub-task
> Reporter: Andrew Purtell
>
> Variation on webtable.
> Instead of crawler or crawler simulation for generating content, configure
> log4j to use an appender that writes into HBase. Optionally run at DEBUG log
> level to increase the evil.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
