[
https://issues.apache.org/jira/browse/HDFS-347?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12764603#action_12764603
]
Todd Lipcon commented on HDFS-347:
----------------------------------
Hey Dhruba,
The connect-back is definitely still up for discussion. I think it's good from
a security standpoint to verify that the client is speaking to the datanode and
not an imposter. This is definitely the simplest part of the code, though, so
we can easily change it if people disagree with me.
I'm still trying to figure out the reason for the overhead. So far, my thoughts
are:
# Checksumming (I was comparing to RawLocalFileSystem, not ChecksumFileSystem).
This is better in 0.21 with the new PureJavaCrc32, but still accounts for some
overhead
# In the above measurements I'm using FileChannel.map to get MappedByteBuffers
for the block and metadata files, then using .get() to do copies into the
provided arrays. Profiling shows most of the time in
java.nio.Bits.copyToByteArray. Right now all transfers from these mapped
buffers are checksum-sized (512 bytes by default) and there appears to be a lot
of overhead there. Next order of business, performance wise, is to see if
introducing a 64KB byte[] buffer will improve things somewhat. This does not
apply to BlockSender, though, since that already forms packets of (I think) 10
checksum chunks at a time.
More theories of course are welcome :)
http://nadeausoftware.com/articles/2008/02/java_tip_how_read_files_quickly is
an interesting resource on this topic as well.
> DFS read performance suboptimal when client co-located on nodes with data
> -------------------------------------------------------------------------
>
> Key: HDFS-347
> URL: https://issues.apache.org/jira/browse/HDFS-347
> Project: Hadoop HDFS
> Issue Type: Improvement
> Reporter: George Porter
> Attachments: HADOOP-4801.1.patch, HADOOP-4801.2.patch,
> HADOOP-4801.3.patch, local-reads-doc
>
>
> One of the major strategies Hadoop uses to get scalable data processing is to
> move the code to the data. However, putting the DFS client on the same
> physical node as the data blocks it acts on doesn't improve read performance
> as much as expected.
> After looking at Hadoop and O/S traces (via HADOOP-4049), I think the problem
> is due to the HDFS streaming protocol causing many more read I/O operations
> (iops) than necessary. Consider the case of a DFSClient fetching a 64 MB
> disk block from the DataNode process (running in a separate JVM) running on
> the same machine. The DataNode will satisfy the single disk block request by
> sending data back to the HDFS client in 64-KB chunks. In BlockSender.java,
> this is done in the sendChunk() method, relying on Java's transferTo()
> method. Depending on the host O/S and JVM implementation, transferTo() is
> implemented as either a sendfilev() syscall or a pair of mmap() and write().
> In either case, each chunk is read from the disk by issuing a separate I/O
> operation for each chunk. The result is that the single request for a 64-MB
> block ends up hitting the disk as over a thousand smaller requests for 64-KB
> each.
> Since the DFSClient runs in a different JVM and process than the DataNode,
> shuttling data from the disk to the DFSClient also results in context
> switches each time network packets get sent (in this case, the 64-kb chunk
> turns into a large number of 1500 byte packet send operations). Thus we see
> a large number of context switches for each block send operation.
> I'd like to get some feedback on the best way to address this, but I think
> providing a mechanism for a DFSClient to directly open data blocks that
> happen to be on the same machine. It could do this by examining the set of
> LocatedBlocks returned by the NameNode, marking those that should be resident
> on the local host. Since the DataNode and DFSClient (probably) share the
> same hadoop configuration, the DFSClient should be able to find the files
> holding the block data, and it could directly open them and send data back to
> the client. This would avoid the context switches imposed by the network
> layer, and would allow for much larger read buffers than 64KB, which should
> reduce the number of iops imposed by each read block operation.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
