Uploaded image for project: 'Lustre'
  1. Lustre
  2. LU-744

Single client's performance degradation on 2.1

Details

    • Bug
    • Resolution: Duplicate
    • Critical
    • None
    • Lustre 2.2.0, Lustre 2.3.0
    • None
    • 3
    • 4018

    Description

      During the performance testing on lustre-2.1, I saw the single client's performance degradation on it.
      Here is IOR results on the single cleints with 2.1 and also lustre-1.8.6.80 for comparing.
      I ran IOR (IOR -t 1m -b 32g -w -r -vv -F -o /lustre/ior.out/file) on the single client with 1, 2, 4 and 8 processes.

      Write(MiB/sec)
      v1.8.6.80 v2.1
      446.25 411.43
      808.53 761.30
      1484.18 1151.41
      1967.42 1172.06

      Read(MiB/sec)
      v1.8.6.80 v2.1
      823.90 595.71
      1449.49 1071.76
      2502.49 1517.79
      3133.43 1746.30

      Tested on same infrastracture(hardware and network). The client just turned off the checksum on both testing.

      Attachments

        1. 2.4 Single Client 3May2013.xlsx
          34 kB
        2. 574.1.pdf
          169 kB
        3. ior-256gb.tar.gz
          32 kB
        4. ior-32gb.tar.gz
          24 kB
        5. lu744-20120909.tar.gz
          883 kB
        6. lu744-20120915.tar.gz
          874 kB
        7. lu744-20120915-02.tar.gz
          1.02 MB
        8. lu744-20121111.tar.gz
          849 kB
        9. lu744-20121113.tar.gz
          846 kB
        10. lu744-20121117.tar.gz
          2.45 MB
        11. lu744-20130104.tar.gz
          915 kB
        12. lu744-20130104-02.tar.gz
          26 kB
        13. lu744-dls-20121113.tar.gz
          10 kB
        14. orig-collectl.out
          81 kB
        15. orig-ior.out
          2 kB
        16. orig-opreport-l.out
          146 kB
        17. patched-collectl.out
          34 kB
        18. patched-ior.out
          2 kB
        19. patched-opreport-l.out
          137 kB
        20. single-client-performance.xlsx
          42 kB
        21. stats-1.8.zip
          14 kB
        22. stats-2.1.zip
          64 kB
        23. test2-various-version.zip
          264 kB
        24. test-patchset-2.zip
          147 kB

        Issue Links

          Activity

            [LU-744] Single client's performance degradation on 2.1
            pjones Peter Jones added a comment -

            This should have been addressed by LU-3321

            pjones Peter Jones added a comment - This should have been addressed by LU-3321
            cliffw Cliff White (Inactive) added a comment - - edited

            Servers are Intel Xeon, 64GB RAM. IOR options were taken from this bug, -t 1m -b 32g
            Client had same/similar CPU, 64GB RAM

            cliffw Cliff White (Inactive) added a comment - - edited Servers are Intel Xeon, 64GB RAM. IOR options were taken from this bug, -t 1m -b 32g Client had same/similar CPU, 64GB RAM

            Cliff, what are server's CPU type, memory size? IOR options, file size. The performance depends client's specs, network and storage.
            We are getting much better perforamnce on the current master and 1.8 client is still fast on some of numbers.
            I will file the latest numbers here.

            ihara Shuichi Ihara (Inactive) added a comment - Cliff, what are server's CPU type, memory size? IOR options, file size. The performance depends client's specs, network and storage. We are getting much better perforamnce on the current master and 1.8 client is still fast on some of numbers. I will file the latest numbers here.

            Test single client performance against 2.3.64 servers, versions tested: 1.8.8, 2.1.5,2.3.0,2.3.64

            cliffw Cliff White (Inactive) added a comment - Test single client performance against 2.3.64 servers, versions tested: 1.8.8, 2.1.5,2.3.0,2.3.64

            All patches have been landed. More work is also needed.

            jay Jinshan Xiong (Inactive) added a comment - All patches have been landed. More work is also needed.

            Andreas,

            As far as I tested, 4943 helped perforamnce improveemnts, but even that patches applied, perforamnce is still lower than b1_8.

            ihara Shuichi Ihara (Inactive) added a comment - Andreas, As far as I tested, 4943 helped perforamnce improveemnts, but even that patches applied, perforamnce is still lower than b1_8.

            Jinshan,
            with http://review.whamcloud.com/4943 landed to master, are there any patches left to land under this bug, or can it be closed?

            adilger Andreas Dilger added a comment - Jinshan, with http://review.whamcloud.com/4943 landed to master, are there any patches left to land under this bug, or can it be closed?

            Gregoire, that's interesting. I wouldn't immediately expect #2929 to make much of a performance impact. How many iterations did you run? I'm curious if those numbers are within the natural variance of the test, or if they're actually because of the changes in #2929. Jinshan, would you expect performance to increase because of that patch?

            prakash Prakash Surya (Inactive) added a comment - Gregoire, that's interesting. I wouldn't immediately expect #2929 to make much of a performance impact. How many iterations did you run? I'm curious if those numbers are within the natural variance of the test, or if they're actually because of the changes in #2929. Jinshan, would you expect performance to increase because of that patch?

            Jinshan,

            What is the status of the patch http://review.whamcloud.com/#change,2929 you posted several months ago for b2_1 release ?
            Why has it never been landed ?

            I have made some measurements and results are significant: from 4% to 50% improvement depending on the platform I tested on.

            Here are the results.

            Hardware configuration:
            30 OSTs
            2 OSS : 4 sockets, 32 cores, 64GB memory, 2xIB, 4xFC8-2port
            ClientA : 4 sockets Nehalem-EX, 32 cores, 64GB memory, 1xIB
            ClientB : 2 sockets SandyBridge-EP, 16 cores, 64GB memory, 1xIB
            Interconnect is QDR Infiniband

            Software configuration:
            kernel 2.6.32-220
            lustre 2.1.3 + ORNL-22 + a few other patches

            IOR file per process, blockSize=4GiB, xfersize=1MiB, fsync=1.
            This gives an aggregate filesize of 120 GiB.

                      #tasks    write   read   configuration
            ClientA       30     1121   1079   lustre 2.1.3
            ClientA       30     1782   1413   lustre 2.1.3 + #2929
            
            ClientB       16     2482   2149   lustre 2.1.3
            ClientB       16     2616   2244   lustre 2.1.3 + #2929
            
            pichong Gregoire Pichon added a comment - Jinshan, What is the status of the patch http://review.whamcloud.com/#change,2929 you posted several months ago for b2_1 release ? Why has it never been landed ? I have made some measurements and results are significant: from 4% to 50% improvement depending on the platform I tested on. Here are the results. Hardware configuration: 30 OSTs 2 OSS : 4 sockets, 32 cores, 64GB memory, 2xIB, 4xFC8-2port ClientA : 4 sockets Nehalem-EX, 32 cores, 64GB memory, 1xIB ClientB : 2 sockets SandyBridge-EP, 16 cores, 64GB memory, 1xIB Interconnect is QDR Infiniband Software configuration: kernel 2.6.32-220 lustre 2.1.3 + ORNL-22 + a few other patches IOR file per process, blockSize=4GiB, xfersize=1MiB, fsync=1. This gives an aggregate filesize of 120 GiB. #tasks write read configuration ClientA 30 1121 1079 lustre 2.1.3 ClientA 30 1782 1413 lustre 2.1.3 + #2929 ClientB 16 2482 2149 lustre 2.1.3 ClientB 16 2616 2244 lustre 2.1.3 + #2929

            new test results includes b1_8, master and master+patch.

            ihara Shuichi Ihara (Inactive) added a comment - new test results includes b1_8, master and master+patch.

            People

              jay Jinshan Xiong (Inactive)
              ihara Shuichi Ihara (Inactive)
              Votes:
              1 Vote for this issue
              Watchers:
              35 Start watching this issue

              Dates

                Created:
                Updated:
                Resolved: