[
https://issues.apache.org/jira/browse/HDDS-6818?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17544356#comment-17544356
]
Stephen O'Donnell commented on HDDS-6818:
-----------------------------------------
{quote}
Due to the incremental requirement of business and the shortage of server slots
in data centers, more and more companies are taking advantage of the deployment
of multi data center.
{quote}
I thought the main point of multi-datacenter is for disaster recovery, rather
than space issues in a data centre? I think in a multi-DC setup, we should be
aiming for a solution where all data is in both DCs. Ideally immediately when
the data is committed, or at least a short while later. When all data is in
both DCs, reads are a fairly simple problem - the nodes come back ordered by
distances from the client and it tries to use the closest one.
Writes are the difficult part, and I don't really see any solution to that in
the design here, aside from keeping all the data in the same DC, which I think
is too big a constraint to put on the design, as it greatly limits its uses.
> Multi DC feature for Ozone
> --------------------------
>
> Key: HDDS-6818
> URL: https://issues.apache.org/jira/browse/HDDS-6818
> Project: Apache Ozone
> Issue Type: New Feature
> Reporter: Janus Chow
> Assignee: Janus Chow
> Priority: Major
>
> The aim of this Jira is to add the multi DC feature for Ozone. This feature
> will try to solve the latency and bandwidth issues of multi DC in two phases.
> Attached the design doc for reviewing:
> https://docs.google.com/document/d/1J_HTIQJ7x8BIMIaSuJPDhaJBLc4WdOqFJUCvCbD3MLE/edit?usp=sharing
--
This message was sent by Atlassian Jira
(v8.20.7#820007)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscr...@ozone.apache.org
For additional commands, e-mail: issues-h...@ozone.apache.org
