Results 1 to 4 of 4

Thread: Kernel 3.0.0 disk io performance drop

  1. #1
    Join Date
    Jul 2011
    Posts
    7

    Default Kernel 3.0.0 disk io performance drop

    Using lvm with two disks I have experienced a significant disk IO performance drop on 3.0.0
    compared to 2.6.39.3 more significantly on random write IO.

    Has someone else experienced similar?


    Kernel 2.6.39.3
    Documents/iobenches/2.6.39.3_512M_2G
    Iozone: Performance Test of File I/O
    Version $Revision: 3.373 $
    Compiled for 64 bit mode.
    Build: linux-AMD64

    Contributors:William Norcott, Don Capps, Isom Crawford, Kirby Collins
    Al Slater, Scott Rhine, Mike Wisner, Ken Goss
    Steve Landherr, Brad Smith, Mark Kelly, Dr. Alain CYR,
    Randy Dunlap, Mark Montague, Dan Million, Gavin Brebner,
    Jean-Marc Zucconi, Jeff Blomberg, Benny Halevy, Dave Boone,
    Erik Habbinga, Kris Strecker, Walter Wong, Joshua Root,
    Fabrice Bacchella, Zhenghua Xue, Qin Li, Darren Sawyer.

    Run began: Mon Jul 25 22:51:29 2011

    Auto Mode
    Include close in write timing
    Using Minimum Record Size 1024 KB
    Using Maximum Record Size 1024 KB
    Using minimum file size of 524288 kilobytes.
    Using maximum file size of 2097152 kilobytes.
    Command line used: iozone -a -c -y1m -q1m -n512M -g2G -L64 -S3072 -i0 -i2
    Output is in Kbytes/sec
    Time Resolution = 0.000001 seconds.
    Processor cache size set to 3072 Kbytes.
    Processor cache line size set to 64 bytes.
    File stride size set to 17 * record size.
    ___________________________ random random bkwd record stride
    KB reclen write rewrite read reread read write read rewrite read fwrite frewrite fread freread
    524288 1024 1010151 1552456 4246417 1783257
    1048576 1024 653349 1476615 3965058 832824
    2097152 1024 345539 357652 3865020 193520

    iozone test complete.
    In Seconds: 43

    Kernel 3.0.0
    Documents/iobenches/3.0.0_512M_2G
    Iozone: Performance Test of File I/O
    Version $Revision: 3.373 $
    Compiled for 64 bit mode.
    Build: linux-AMD64

    Contributors:William Norcott, Don Capps, Isom Crawford, Kirby Collins
    Al Slater, Scott Rhine, Mike Wisner, Ken Goss
    Steve Landherr, Brad Smith, Mark Kelly, Dr. Alain CYR,
    Randy Dunlap, Mark Montague, Dan Million, Gavin Brebner,
    Jean-Marc Zucconi, Jeff Blomberg, Benny Halevy, Dave Boone,
    Erik Habbinga, Kris Strecker, Walter Wong, Joshua Root,
    Fabrice Bacchella, Zhenghua Xue, Qin Li, Darren Sawyer.

    Run began: Mon Jul 25 22:57:06 2011

    Auto Mode
    Include close in write timing
    Using Minimum Record Size 1024 KB
    Using Maximum Record Size 1024 KB
    Using minimum file size of 524288 kilobytes.
    Using maximum file size of 2097152 kilobytes.
    Command line used: iozone -a -c -y1m -q1m -n512M -g2G -L64 -S3072 -i0 -i2
    Output is in Kbytes/sec
    Time Resolution = 0.000001 seconds.
    Processor cache size set to 3072 Kbytes.
    Processor cache line size set to 64 bytes.
    File stride size set to 17 * record size.
    __________________________ random random bkwd record stride
    KB reclen write rewrite read reread read write read rewrite read fwrite frewrite fread freread
    524288 1024 1017221 1574878 4211487 1817301
    1048576 1024 515143 1494402 3842119 212495
    2097152 1024 352130 334563 4030001 167668

    iozone test complete.
    In Seconds: 48

    Kernel 2.6.39.3
    Documents/iobenches/fio_26393
    Flexible IO Tester:
    pts/fio-1.1.0 [Test: Intel IOMeter File Server Access Pattern]
    Test 1 of 1
    Estimated Time Remaining: 3 Minutes
    Expected Trial Run Count: 3
    Running Pre-Test Script @ 21:12:13
    Started Run 1 @ 21:12:13
    Running Interim-Test Script @ 21:12:40
    Started Run 2 @ 21:12:42
    Running Interim-Test Script @ 21:13:06
    Started Run 3 @ 21:13:09 [Std. Dev: 1.26%]
    Running Post-Test Script @ 21:13:33

    Test Results:
    23.610690116882
    24.161954879761
    24.096673965454

    Average: 23.96 Seconds (Run Time)


    Kernel 3.0.0
    Documents/iobenches/fio_300
    Estimated Run-Time: 2 Minutes



    Flexible IO Tester:
    pts/fio-1.1.0 [Test: Intel IOMeter File Server Access Pattern]
    Test 1 of 1
    Estimated Time Remaining: 2 Minutes
    Expected Trial Run Count: 3
    Running Pre-Test Script @ 21:02:48
    Started Run 1 @ 21:02:48
    Running Interim-Test Script @ 21:03:22
    Started Run 2 @ 21:03:24
    Running Interim-Test Script @ 21:04:24
    Started Run 3 @ 21:04:26
    Running Interim-Test Script @ 21:04:53 [Std. Dev: 46.97%]
    Started Run 4 @ 21:04:55
    Running Interim-Test Script @ 21:05:52 [Std. Dev: 40.18%]
    Started Run 5 @ 21:05:54
    Running Interim-Test Script @ 21:06:20 [Std. Dev: 42.57%]
    Started Run 6 @ 21:06:22 [Std. Dev: 39.08%]
    Running Post-Test Script @ 21:07:19

    Test Results:
    30.38679599762
    60.021185159683
    26.55930685997
    57.128906011581
    26.059051990509
    57.044394016266

    Average: 42.87 Seconds (Run Time)

  2. #2
    Join Date
    Jul 2011
    Posts
    7

    Default Memory / disk performance drop


  3. #3
    Join Date
    Jul 2008
    Location
    Greece
    Posts
    3,788

    Default

    Quote Originally Posted by sforum View Post
    No, "404 - Not Found" is definitely not the cause of this.

  4. #4
    Join Date
    Jul 2011
    Posts
    7

    Default Link to kernel disk performance drop


Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •