You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@hbase.apache.org by "stack (JIRA)" <ji...@apache.org> on 2018/05/22 21:36:00 UTC

[jira] [Updated] (HBASE-20620) HBASE-20564 Tighter ByteBufferKeyValue Cell Comparator; part 2

     [ https://issues.apache.org/jira/browse/HBASE-20620?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

stack updated HBASE-20620:
--------------------------
    Summary: HBASE-20564 Tighter ByteBufferKeyValue Cell Comparator; part 2  (was: HBASE-20564 Tighter ByteBufferKeyValue Cell Comparator part 2)

> HBASE-20564 Tighter ByteBufferKeyValue Cell Comparator; part 2
> --------------------------------------------------------------
>
>                 Key: HBASE-20620
>                 URL: https://issues.apache.org/jira/browse/HBASE-20620
>             Project: HBase
>          Issue Type: Sub-task
>          Components: Performance
>    Affects Versions: 2.0.0
>            Reporter: stack
>            Assignee: stack
>            Priority: Major
>             Fix For: 2.0.1
>
>
> This is a follow-on from HBASE-20564 "Tighter ByteBufferKeyValue Cell Comparator". In this issue, we make a stripped-down comparator that we deploy in one location only, as memstore Comparator. HBASE-20564 operated on CellComparator/Impl and got us 5-10k more throughput on top of a baseline 40k or so. A purposed stripped-down ByteBufferKeyValue comparator that fosters better inlining gets us from 45-50k up to about 75k (1.4 does 110-115k no-WAL PE writes). Data coming. Log of profiling kept here: https://docs.google.com/document/d/1vZ_k6_pNR1eQxID5u1xFihuPC7FkPaJQW8c4M5eA2AQ/edit#



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)