You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@atlas.apache.org by "Hemanth Yamijala (JIRA)" <ji...@apache.org> on 2015/11/30 06:13:11 UTC

[jira] [Created] (ATLAS-341) Doc: Fault tolerant configurations for Atlas backing stores

Hemanth Yamijala created ATLAS-341:
--------------------------------------

             Summary: Doc: Fault tolerant configurations for Atlas backing stores
                 Key: ATLAS-341
                 URL: https://issues.apache.org/jira/browse/ATLAS-341
             Project: Atlas
          Issue Type: Bug
            Reporter: Hemanth Yamijala


Atlas configuration documentation is lacking configuration specific to make backing stores HA.

For e.g.:
* There should be pointers to set up HA for HBase master
* The sample scripts to create the Solr indexes should specifically mention that the number of replicas should be more than one for ensuring availability of indices for search in the face of Solr node failure.
* We create Kafka topics automatically for communicating events from Hive hook (topic ATLAS_HOOK) and also to communicate events out from Atlas (topic ATLAS_ENTITIES). By default these are created without replicas, which means we would lose events if Atlas broker node fails.



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