hbase-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "stack (JIRA)" <j...@apache.org>
Subject [jira] Updated: (HBASE-1393) Narrow synchronization in HLog
Date Fri, 08 May 2009 16:53:45 GMT

     [ https://issues.apache.org/jira/browse/HBASE-1393?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel

stack updated HBASE-1393:

    Attachment: 1393.patch

Remove the HLog.isDeleted and the 'special' HLogEdit.DELETED_BYTES.  We don't do deletes this
way anymore.

Added a sleep to regionserver of 1 second on its main loop if is in the exceeded maximum retries
state trying to send report to master.

Changed the datastructures used by HLog.  Use concurrentskiplistmap instead of using a synchronizedSortedMap
TreeMap.  Should be faster.  Changed all counters to be Atomics.  Volatile counters don't
work: http://pitfalls.wordpress.com/2008/05/25/javavolatile/.  Got rid of sequenceLock.  We
should be able to do without it (We might spin a while setting it on region open is the price
we pay -- in testing haven't seen it).   Narrowed the synchronized section in HLog#append
though I don't think it'll make much difference in scheme of things -- i/o is so big compared.

> Narrow synchronization in HLog
> ------------------------------
>                 Key: HBASE-1393
>                 URL: https://issues.apache.org/jira/browse/HBASE-1393
>             Project: Hadoop HBase
>          Issue Type: Improvement
>         Environment: Prompted by HBASE-1213, review HLog synchronizations.
>            Reporter: stack
>             Fix For: 0.20.0
>         Attachments: 1393.patch

This message is automatically generated by JIRA.
You can reply to this email to add a comment to the issue online.

View raw message