[
https://issues.apache.org/jira/browse/HADOOP-3288?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12591206#action_12591206
]
stack commented on HADOOP-3288:
-------------------------------
> For an application like HBase, what is more important : over all throughput
> (transactions per sec) or latency of serial requests?
If you put the question that way, overall throughput is more important.
Raghu, is the argument that we can't have it both ways?
Making a guess, I'd say that the character of reading on HBase clusters will be
predominantly random reads with some much smaller number of concurrent scans of
all or parts of tables (a 'scan' implies serial reading of files). In the
HBase case, we go out of our way to keep files in HDFS small, never > 256M or
so (about 2-4 or 5 blocks).
> Serial streaming performance should be Math.min(ideal client performance,
> ideal serial hdfs performance)
> --------------------------------------------------------------------------------------------------------
>
> Key: HADOOP-3288
> URL: https://issues.apache.org/jira/browse/HADOOP-3288
> Project: Hadoop Core
> Issue Type: Improvement
> Components: dfs
> Affects Versions: 0.16.3, 0.18.0
> Environment: Mac OS X 10.5.2, Java 6
> Reporter: Sam Pullara
> Fix For: 0.18.0
>
>
> I looked at all the code long and hard and this was my analysis (could be
> wrong, I'm not an expert on this codebase):
> Current Serial HDFS performance = Average Datanode Performance
> Average Datanode Performance = Average Disk Performance (even if you have
> more than one)
> We should have:
> Ideal Serial HDFS Performance = Sum of Ideal Datanode Performance
> Ideal Datanode Performance = Sum of disk performance
> When you read a single file serially from HDFS there are a number of
> limitations that come into play:
> 1) Blocks on multiple datanodes will be load balanced between them -
> averaging the performance of the datanodes
> 2) Blocks on multiple disks in a single datanode are load balanced between
> them - averaging the performance of the disks
> I think that all this could be fixed if we actually prefetched fully read
> blocks on the client until the client can no longer keep up with the data or
> there is another bottleneck like network bandwidth.
> This seems like a reasonably common use case though not the typical MapReduce
> case.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
