[
https://issues.apache.org/jira/browse/HBASE-26890?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17512733#comment-17512733
]
Duo Zhang commented on HBASE-26890:
-----------------------------------
This is related to the WAL refactor issue. We should build sync replication
upon the WAL interface, otherwise once we want to change the WAL
implementation, sync replication will be broken.
Of course there will be new issues if we want to build sync replication upon
WAL interface, for example, now we can make sure that the files are always the
same for local and remote WALs, as they share the same WAL instance and roll at
the same time, but if they are two WAL instances, it is not easy to archive
this. So when deleting remote WAL files when replicating, we need to find a way
to determine whether a file can be deleted. It is also related to how to track
the offset of a replication queue.
Anyway, they are internal implementations, so we are free to change them at any
time, so this is a not a blocker for our 3.0.0 release.
> Make the WAL interface async so sync replication can be built up on the WAL
> interface
> -------------------------------------------------------------------------------------
>
> Key: HBASE-26890
> URL: https://issues.apache.org/jira/browse/HBASE-26890
> Project: HBase
> Issue Type: Improvement
> Reporter: Duo Zhang
> Priority: Major
>
> Instead of hacking into the WAL implementation.
> This could make the implementation more general if later we want to change
> the WAL implementation.
--
This message was sent by Atlassian Jira
(v8.20.1#820001)
