[
https://issues.apache.org/jira/browse/HBASE-20329?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16426410#comment-16426410
]
Anoop Sam John commented on HBASE-20329:
----------------------------------------
That Q was not really on the new WAL impl 'AsyncFSWAL'. That was more on our
Durability semantics ASYNC_WAL. This can be used with whatever WAL impl under
the server. When it is ASYNC_WAL durability, we will not call sync to
complete the write ops. The sync will happen in an async way later. So ya
when crash is there, we may get the data loss. But with our new WAL impl,
AsyncFSWAL there can NOT be data loss. :-).. As this is clubbed with this
AsyncFSWAL doc jira , am saying boss..
Now these names confuses us a lot.
> Add note for operators to refguide on AsyncFSWAL
> ------------------------------------------------
>
> Key: HBASE-20329
> URL: https://issues.apache.org/jira/browse/HBASE-20329
> Project: HBase
> Issue Type: Sub-task
> Components: documentation, wal
> Reporter: stack
> Assignee: stack
> Priority: Critical
> Fix For: 3.0.0
>
> Attachments: HBASE-20329.master.001.patch
>
>
> Need a few notes in refguide on this new facility.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
