[ https://issues.apache.org/jira/browse/HBASE-9996?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13826448#comment-13826448
]
Jean-Marc Spaggiari commented on HBASE-9996:
--------------------------------------------
sgtm
Any risks that this.fs.getFileStatus(oldFile).getLen() changes between the sync and when it's
called just after?
+1
> move the log io operation out of updateLock scope
> -------------------------------------------------
>
> Key: HBASE-9996
> URL: https://issues.apache.org/jira/browse/HBASE-9996
> Project: HBase
> Issue Type: Improvement
> Components: regionserver
> Affects Versions: 0.94.13
> Reporter: Liang Xie
> Assignee: Liang Xie
> Attachments: HBase-9996-0.94.txt
>
>
> I am working at another write outlier issue, and just saw this code pattern:
> {code}
> synchronized(updateLock) {
> ...
> LOG.info(xxx)
> ...
> }
> {code}
> let's move the log statement out of this sync block. seems it's ok in trunk code, just
a 0.94 issue
--
This message was sent by Atlassian JIRA
(v6.1#6144)
|