[
https://issues.apache.org/jira/browse/ATLAS-3057?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Ashutosh Mestry updated ATLAS-3057:
-----------------------------------
Description:
*Background*
For Atlas version that uses _HBase_ and _Solr_, occasionally if Solr is down
and entities get created, then there is no record of the created entities
within Solr. Basic search does not indicate presence of such entities.
*Solution*
Index Repair Tool (which was present in branch-0.8) should be implemented for
JanusGraph.
*Implementation Guidance*
* Create Java-based implementation rather than Groovy script that needs
graphdb-specific shell to be installed.
* Use JanusGraph APIs for restore instead of custom per-vertex logic.
* Investigate possibility of using MapReduce for higher throughput.
was:
*Background*
For Atlas version that uses _HBase_ and _Solr_, occasionally if Solr is down
and entities get created, then there is no record of the created entities
within Solr. Basic search does not indicate presence of such entities.
*Solution*
Index Repair Tool (which was present in branch-0.8) should be implemented for
JanusGraph.
*Implementation Guidance*
* Create Java-based implementation rather than Groovy script that needs
graphdb-specific shell to be installed.
* Use JanusGraph APIs for restore instead of custom per-vertex logic.
> Index Repair Tool: Add JanusGraph-Specific Index Repair Tool
> ------------------------------------------------------------
>
> Key: ATLAS-3057
> URL: https://issues.apache.org/jira/browse/ATLAS-3057
> Project: Atlas
> Issue Type: Bug
> Reporter: Ashutosh Mestry
> Assignee: Ashutosh Mestry
> Priority: Major
>
> *Background*
> For Atlas version that uses _HBase_ and _Solr_, occasionally if Solr is down
> and entities get created, then there is no record of the created entities
> within Solr. Basic search does not indicate presence of such entities.
> *Solution*
> Index Repair Tool (which was present in branch-0.8) should be implemented for
> JanusGraph.
> *Implementation Guidance*
> * Create Java-based implementation rather than Groovy script that needs
> graphdb-specific shell to be installed.
> * Use JanusGraph APIs for restore instead of custom per-vertex logic.
> * Investigate possibility of using MapReduce for higher throughput.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
