[
https://issues.apache.org/jira/browse/HBASE-29228?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Rahul Agarkar updated HBASE-29228:
----------------------------------
Status: Patch Available (was: Open)
> Add support to prevent running multiple active clusters
> -------------------------------------------------------
>
> Key: HBASE-29228
> URL: https://issues.apache.org/jira/browse/HBASE-29228
> Project: HBase
> Issue Type: Sub-task
> Components: master, regionserver
> Reporter: Abhishek Kothalikar
> Assignee: Abhishek Kothalikar
> Priority: Major
> Labels: pull-request-available
>
> Using HBase Active and Read Replica Clusters, there could be potential
> problem scenarios where multiple clusters started in the active mode. For
> example,
> |Active Cluster
> *A*
> hbase.global.readonly.enabled set to false
> Replica Clusters
> *B*
> hbase.global.readonly.enabled set to true
> *C*
> hbase.global.readonly.enabled set to true|
>
> In the above mentioned setup, if
> * the user accidentally sets hbase.global.readonly.enabled to false on
> cluster B
> * user attempts restart on B then, B will start itself as an active cluster
> * moreover, at this point, it cannot be deduced that which cluster is
> ‘original’ active cluster
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
