[
https://issues.apache.org/jira/browse/HDFS-14617?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16880684#comment-16880684
]
Stephen O'Donnell commented on HDFS-14617:
------------------------------------------
The time taken to read and parse the protobuf messages seems to dominate the
runtime, at least for the inode section, so I am not sure a single thread
reading and decoding the data from disk would give the same speed benefits.
Most of the complexity in this patch is to allow loading the inodeDirectory
section in parallel, as it has some computationally expensive parts and needed
a bit of a refactor to allow things to work well in parallel.
I also like that the code to do this in parallel, aside from some
synchronization, is largely the same as the single threaded version, so its
fairly easy to understand what it is doing if you understand the origional
code. I guess a producer / consumer model can be easy to understand too, but we
will need to somehow read and decode the protobuf in parallel to get
significant speedup and work some synchronization into the code for the
consumers to run in parallel if needed too.
These extra section markers would not cause a namenode without this patch to
fail to load it, so the focus would be on external tools that may load the
image for other purposes. How concerned do we need to be about external tools
loading the image for a change like this?
> Improve fsimage load time by writing sub-sections to the fsimage index
> ----------------------------------------------------------------------
>
> Key: HDFS-14617
> URL: https://issues.apache.org/jira/browse/HDFS-14617
> Project: Hadoop HDFS
> Issue Type: Improvement
> Components: namenode
> Reporter: Stephen O'Donnell
> Assignee: Stephen O'Donnell
> Priority: Major
> Attachments: HDFS-14617.001.patch
>
>
> Loading an fsimage is basically a single threaded process. The current
> fsimage is written out in sections, eg iNode, iNode_Directory, Snapshots,
> Snapshot_Diff etc. Then at the end of the file, an index is written that
> contains the offset and length of each section. The image loader code uses
> this index to initialize an input stream to read and process each section. It
> is important that one section is fully loaded before another is started, as
> the next section depends on the results of the previous one.
> What I would like to propose is the following:
> 1. When writing the image, we can optionally output sub_sections to the
> index. That way, a given section would effectively be split into several
> sections, eg:
> {code:java}
> inode_section offset 10 length 1000
> inode_sub_section offset 10 length 500
> inode_sub_section offset 510 length 500
>
> inode_dir_section offset 1010 length 1000
> inode_dir_sub_section offset 1010 length 500
> inode_dir_sub_section offset 1010 length 500
> {code}
> Here you can see we still have the original section index, but then we also
> have sub-section entries that cover the entire section. Then a processor can
> either read the full section in serial, or read each sub-section in parallel.
> 2. In the Image Writer code, we should set a target number of sub-sections,
> and then based on the total inodes in memory, it will create that many
> sub-sections per major image section. I think the only sections worth doing
> this for are inode, inode_reference, inode_dir and snapshot_diff. All others
> tend to be fairly small in practice.
> 3. If there are under some threshold of inodes (eg 10M) then don't bother
> with the sub-sections as a serial load only takes a few seconds at that scale.
> 4. The image loading code can then have a switch to enable 'parallel loading'
> and a 'number of threads' where it uses the sub-sections, or if not enabled
> falls back to the existing logic to read the entire section in serial.
> Working with a large image of 316M inodes and 35GB on disk, I have a proof of
> concept of this change working, allowing just inode and inode_dir to be
> loaded in parallel, but I believe inode_reference and snapshot_diff can be
> make parallel with the same technique.
> Some benchmarks I have are as follows:
> {code:java}
> Threads 1 2 3 4
> --------------------------------
> inodes 448 290 226 189
> inode_dir 326 211 170 161
> Total 927 651 535 488 (MD5 calculation about 100 seconds)
> {code}
> The above table shows the time in seconds to load the inode section and the
> inode_directory section, and then the total load time of the image.
> With 4 threads using the above technique, we are able to better than half the
> load time of the two sections. With the patch in HDFS-13694 it would take a
> further 100 seconds off the run time, going from 927 seconds to 388, which is
> a significant improvement. Adding more threads beyond 4 has diminishing
> returns as there are some synchronized points in the loading code to protect
> the in memory structures.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: hdfs-issues-unsubscr...@hadoop.apache.org
For additional commands, e-mail: hdfs-issues-h...@hadoop.apache.org
