[
https://issues.apache.org/jira/browse/HADOOP-10206?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Dexter Bradshaw updated HADOOP-10206:
-------------------------------------
Changes are being reviewed on the HWX JIRA at
https://hwxmonarch.atlassian.net/browse/HADOOP. Two outstanding items remain to
be completed:
a. TODO: Perform internal hsync on a smaller bounded queue when it approaches
capacity approaching capacity.
b. TODO: The I/O queue is single threaded. Mostafa claims that this is for
correctness. I have to investigate whether we can in fact do overlapped
parallel I/O’s since the addresses of the log blocks are already determined by
the streaming interface.
The temporary fix with the unbounded queue is coded and now works with large
tables in YCSB.
> Port WASB HBase support to Hadoop 2.0
> -------------------------------------
>
> Key: HADOOP-10206
> URL: https://issues.apache.org/jira/browse/HADOOP-10206
> Project: Hadoop Common
> Issue Type: Improvement
> Affects Versions: 2.2.0
> Environment: Windows, Windows Azure
> Reporter: Dexter Bradshaw
> Labels: patch
> Original Estimate: 504h
> Remaining Estimate: 504h
>
> A series of changes for HBase support on Hadoop 2.0. These changes include
> support for page blobs, fixes to allows HBase logging to block blobs, and
> support for hsync and hflush methods in the Syncable interface.
--
This message was sent by Atlassian JIRA
(v6.1.5#6160)
