[
https://issues.apache.org/jira/browse/HBASE-27110?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17702976#comment-17702976
]
Guanghao Zhang commented on HBASE-27110:
----------------------------------------
HBase used hbase:meta or ZK as meta info storage. Then HBase 2.x introduced a
new master local region to store some meta info. Need a discussion about the
long-term road map about meta info storage? It's not just a problem of hbase
replication peer storage.
> Move replication peer storage from zookeeper to other storage systems
> ---------------------------------------------------------------------
>
> Key: HBASE-27110
> URL: https://issues.apache.org/jira/browse/HBASE-27110
> Project: HBase
> Issue Type: New Feature
> Components: Replication
> Reporter: Duo Zhang
> Priority: Major
>
> This is a more specific issue based on the works which are already done in
> HBASE-15867.
> There are several candidates for storing replication peer.
> 1. A new family of master local region
> 2. On FileSystem
> 3. A new family of hbase:meta
> Either choice has pros and cons. We need to decide which way to go first.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
