hbase-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "stack (JIRA)" <j...@apache.org>
Subject [jira] [Resolved] (HBASE-11160) Undo append waiting on region edit/sequence id update
Date Mon, 06 Oct 2014 19:15:34 GMT

     [ https://issues.apache.org/jira/browse/HBASE-11160?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel

stack resolved HBASE-11160.
    Resolution: Later

Resolving as 'later'. We lose some writethrough speed because we wait on sequencedid.  Would
be good to fix but does not look easy.  We can come back to this later post WAL Interfaces
redo for multi-WAL.

> Undo append waiting on region edit/sequence id update
> -----------------------------------------------------
>                 Key: HBASE-11160
>                 URL: https://issues.apache.org/jira/browse/HBASE-11160
>             Project: HBase
>          Issue Type: Sub-task
>          Components: wal
>    Affects Versions: 0.99.0
>            Reporter: stack
>            Assignee: stack
>            Priority: Critical
>             Fix For: 2.0.0, 0.99.1
> HBASE--11135 changed append to wait on the update of the edit/sequence id.  This slows
writes by a significant amount (50% slower?).  [~jeffreyz] is going to look at updating MVCC
inside in the ring buffer consumer thread just after the region edit/sequence id update. 
If that works, then we can undo having append wait on the sequence id update and get some
of our speedup back.  Marking this 0.99 and blocker so we don't forget about it (Undoing the
wait is just a case of removing the few lines of latch code in FSWALEdit).

This message was sent by Atlassian JIRA

View raw message