[ 
https://issues.apache.org/jira/browse/HBASE-18070?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17232859#comment-17232859
 ] 

Andrew Kyle Purtell commented on HBASE-18070:
---------------------------------------------

{quote}And I think here we could do much better than ‘HedgeRead’ solution.
The ‘HedgeRead’ mode will always request to primary replica first, and then 
fallback to secondary replicas when it can not get the result back in time. 
This means all traffic will go to primary first until we make the primary 
replica down, we will go to secondary replicas. But here we will go to 
different replicas evenly, which should have better performance when the 
traffic could make a single replica down.
{quote}
I completely agree with [~zhangduo] on this point. It would be good to choose 
between an "any replica" or "primary first" access pattern. In my opinion we 
could have made that a follow up issue for improvement. I.e. this feature is 
the main work; this alternate mode is a modest increment.

I wish this discussion had taken place closer to when we agreed to hold the 2.4 
RC. We've been waiting for two weeks there for this to merge. If this change 
caused regression or failed its new unit tests of course it would not go in. 
This isn't a blocker like that, more like a disagreement on technical direction 
that could be solved with some follow up. On the other hand we have a (IMHO) 
valid veto on the merge thread. In the interest of unblocking 2.4 let's move 
this out until the next one. By then hopefully everyone is satisfied.

> Enable memstore replication for meta replica
> --------------------------------------------
>
>                 Key: HBASE-18070
>                 URL: https://issues.apache.org/jira/browse/HBASE-18070
>             Project: HBase
>          Issue Type: New Feature
>            Reporter: Hua Xiang
>            Assignee: Huaxiang Sun
>            Priority: Critical
>             Fix For: 3.0.0-alpha-1, HBASE-18070, HBASE-18070.branch-2, 2.5.0
>
>
> Based on the current doc, memstore replication is not enabled for meta 
> replica. Memstore replication will be a good improvement for meta replica. 
> Create jira to track this effort (feasibility, design, implementation, etc).



--
This message was sent by Atlassian Jira
(v8.3.4#803005)

Reply via email to