[jira] [Commented] (HBASE-18906) Investigate Phoenix usages of Region#waitForXXX APIs

["Andrew Purtell (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/HBASE-18906?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16186705#comment-16186705 ] 

Andrew Purtell commented on HBASE-18906:
----------------------------------------

Phoenix doesn't use these APIs, but they should. Instead they are doing this terrible thing: https://github.com/apache/phoenix/blob/master/phoenix-core/src/main/java/org/apache/phoenix/coprocessor/UngroupedAggregateRegionObserver.java#L236 

I think we should consider providing APIs for the following:
- Request flush async
- Request compaction async
- Poll for the completion of a single flush request
- Poll for the completion of a single compaction request
- Wait for all pending flushes to complete on the region
- Wait for all pending compactions to finish on the region

It's a good idea to get away from providing access to low level implementation details. However I think it would be good to allow coprocessors to wrap the behaviors. One example: guarantee persistence in HFiles: write some data, trigger a flush (API needed), wait for flushes to complete (API needed), then do more stuff. Another example: secure delete: lay some delete markers, trigger flushing and compactions (APIs needed), wait for the activity to complete (API needed), then do more stuff. 

> Investigate Phoenix usages of Region#waitForXXX APIs
> ----------------------------------------------------
>
>                 Key: HBASE-18906
>                 URL: https://issues.apache.org/jira/browse/HBASE-18906
>             Project: HBase
>          Issue Type: Sub-task
>          Components: Coprocessors
>            Reporter: Anoop Sam John
>            Assignee: Anoop Sam John
>             Fix For: 2.0.0-alpha-4
>
>
> While reviewing HBASE-18183, Andy pointed out that Phoenix uses waitForFlushesAndCompactions and/or waitForFlushes for diff reasons.  This issue is to see why they need them and whether alternate ways are possible. This seems to be too much internal stuff and a normal CP hook calling these would be dangerous.
> If there are alternate ways for Phoenix not to use this and not landing in issues (As said by Andy) we should suggest/fix for them.



--
This message was sent by Atlassian JIRA
(v6.4.14#64029)