Home | About | Sematext search-lucene.com search-hadoop.com
 Search Hadoop and all its subprojects:

Switch to Threaded View
MapReduce, mail # user - How to test the performance of NN?


Copy link to this message
-
Re: How to test the performance of NN?
Mark Kerzner 2013-06-07, 01:41
Awesome advice, Ivan! It works for me now.

Thank you,
Mark
On Thu, Jun 6, 2013 at 12:18 AM, Ivan Mitic <[EMAIL PROTECTED]> wrote:

>  Hi Mark,****
>
> ** **
>
> Your NNBench output does not look OK to me. Below is an example NNBench
> cmd line I used some time ago and its output.****
>
> ** **
>
> Btw, what I found useful  in testing the disk IO performance for namenode
> operations (others might find this unusual) is the TestEditLog unittest. I
> would modify the test as appropriate and run the simulation. This was a
> quick micro-benchmark I used to compare a couple of devices. Note that
> you’ll have to modify the test in trunk and remove the line with
> EditLogFileOutputStream.setShouldSkipFsyncForTesting(), otherwise you’re
> not testing the real thing.****
>
> ** **
>
> ** **
>
> hadoop.cmd jar hadoop-test-1.1.0-SNAPSHOT.jar nnbench -operation
> create_write -maps 32 -reduces 1 -blockSize 1 -bytesToWrite 20
> -bytesPerChecksum 1 -numberOfFiles 500 -replicationFactorPerFile 1****
>
> ** **
>
> ** **
>
> -------------- NNBench -------------- :****
>
>                                 Version: NameNode Benchmark 0.4****
>
>                             Date & time: 2012-12-18 07:29:20,832****
>
> ** **
>
>                          Test Operation: create_write****
>
>                              Start time: 2012-12-18 07:25:36,111****
>
>                             Maps to run: 32****
>
>                          Reduces to run: 1****
>
>                      Block Size (bytes): 1****
>
>                          Bytes to write: 20****
>
>                      Bytes per checksum: 1****
>
>                         Number of files: 500****
>
>                      Replication factor: 1****
>
>              Successful file operations: 16000****
>
> ** **
>
>          # maps that missed the barrier: 0****
>
>                            # exceptions: 0****
>
> ** **
>
>                 TPS: Create/Write/Close: 153****
>
> Avg exec time (ms): Create/Write/Close: 301.6148125****
>
>              Avg Lat (ms): Create/Write: 68.25625****
>
>                     Avg Lat (ms): Close: 233.245****
>
> ** **
>
>                   RAW DATA: AL Total #1: 1092100****
>
>                   RAW DATA: AL Total #2: 3731920****
>
>                RAW DATA: TPS Total (ms): 4825837****
>
>         RAW DATA: Longest Map Time (ms): 208692.0****
>
>                     RAW DATA: Late maps: 0****
>
>               RAW DATA: # of exceptions: 0****
>
> ** **
>
> Hope this helps,****
>
> Ivan****
>
> ** **
>
> *From:* Mark Kerzner [mailto:[EMAIL PROTECTED]]
> *Sent:* Wednesday, June 05, 2013 5:16 PM
> *To:* Hadoop User
> *Subject:* Re: How to test the performance of NN?****
>
> ** **
>
> Hi, Ivan,****
>
> thank you for your willingness to help. First, what do I mean by a more
> efficient NN? - I am experimenting with Kove <http://kove.com/> device,
> persistent memory. I want to do one of the two things:****
>
> 1. Use it for FSDirectory, with the expectation of very fast writes, to
> see if this improves performance.****
>
> 2. Use it for BlockMaps, not for performance, but for fault-tolerance.****
>
> To prepare for this, I want to run some benchmarks on the given install
> just to get experience running it. Then I will run it on a 3-node cluster
> without Kove, and then with Kove.****
>
> Here is what I got from nnbench, and I am trying to find what I should
> watch for to measure NN performance. Does this all make sense?****
>
> Thank you,
> Mark
>
>
> 13/06/05 16:50:16 INFO hdfs.NNBench: -------------- NNBench --------------
> :
> 13/06/05 16:50:16 INFO hdfs.NNBench:
> Version: NameNode Benchmark 0.4
> 13/06/05 16:50:16 INFO hdfs.NNBench:                            Date &
> time: 2013-06-05 16:50:16,412
> 13/06/05 16:50:16 INFO hdfs.NNBench:
> 13/06/05 16:50:16 INFO hdfs.NNBench:                         Test
> Operation: open_read
> 13/06/05 16:50:16 INFO hdfs.NNBench:                             Start
> time: 2013-06-05 16:48:34,792
> 13/06/05 16:50:16 INFO hdfs.NNBench:                            Maps to