Zhijie Shen commented on YARN-3411:

[~vrushalic], thanks for working on the patch. Some comment from my side:

1. I saw in HBase implementation flow version is not included as part of row 
key. This is a bit different from primary key design of Phoenix implementation. 
Would you mind elaborating your rationale a bit?

2. Shall we make the constants in TimelineEntitySchemaConstants follow Hadoop 
convention? We can keep them in this class now. Once we decide to move on with 
HBase impl, we should move (some of) them into YarnConfiguration as API.

3. I saw the classes are marked \@Public, but they're the backend classes and 
not accessible by the user directly. In fact, you can leave these classes not 

4. According to TimelineSchemaCreator, we need to run command line to create 
the table when we setup the backend, right? Can we include creating the table 
into the lifecycle of HBaseTimelineWriterImpl?

> [Storage implementation] explore the native HBase write schema for storage
> --------------------------------------------------------------------------
>                 Key: YARN-3411
>                 URL: https://issues.apache.org/jira/browse/YARN-3411
>             Project: Hadoop YARN
>          Issue Type: Sub-task
>          Components: timelineserver
>            Reporter: Sangjin Lee
>            Assignee: Vrushali C
>            Priority: Critical
>         Attachments: ATSv2BackendHBaseSchemaproposal.pdf, 
> YARN-3411-YARN-2928.001.patch, YARN-3411-YARN-2928.002.patch, 
> YARN-3411-YARN-2928.003.patch, YARN-3411-YARN-2928.004.patch, 
> YARN-3411-YARN-2928.005.patch, YARN-3411-YARN-2928.006.patch, 
> YARN-3411.poc.2.txt, YARN-3411.poc.3.txt, YARN-3411.poc.4.txt, 
> YARN-3411.poc.5.txt, YARN-3411.poc.6.txt, YARN-3411.poc.7.txt, 
> YARN-3411.poc.txt
> There is work that's in progress to implement the storage based on a Phoenix 
> schema (YARN-3134).
> In parallel, we would like to explore an implementation based on a native 
> HBase schema for the write path. Such a schema does not exclude using 
> Phoenix, especially for reads and offline queries.
> Once we have basic implementations of both options, we could evaluate them in 
> terms of performance, scalability, usability, etc. and make a call.

This message was sent by Atlassian JIRA

Reply via email to