Find the answer to your Linux question:
Results 1 to 4 of 4
Here is a capture of what iostat -d -x 5 3 spit out: Device: rrqm/s wrqm/s r/s w/s rsec/s wsec/s rkB/s wkB/s avgrq-sz avgqu-sz await svctm %util sda 0.25 6.99 ...
Enjoy an ad free experience by logging in. Not a member yet? Register.
  1. #1
    Just Joined!
    Join Date
    Dec 2012
    Posts
    3

    I need help figuring out if my system has high IO usage.


    Here is a capture of what iostat -d -x 5 3 spit out:

    Device: rrqm/s wrqm/s r/s w/s rsec/s wsec/s rkB/s wkB/s avgrq-sz avgqu-sz await svctm %util
    sda 0.25 6.99 2.99 2.29 134.21 76.33 67.10 38.17 39.81 0.92 174.48 55.07 29.13
    sdb 0.00 5.41 11.26 34.96 102.36 117.20 51.18 58.60 4.75 1.00 21.54 10.23 47.26
    sdc 0.00 7.96 24.52 51.14 213.89 462.07 106.95 231.03 8.93 0.59 7.83 9.14 69.14

    Device: rrqm/s wrqm/s r/s w/s rsec/s wsec/s rkB/s wkB/s avgrq-sz avgqu-sz await svctm %util
    sda 0.40 1.00 10.60 6.20 302.40 56.00 151.20 28.00 21.33 0.57 32.71 6.62 11.12
    sdb 0.00 18.80 29.20 79.60 238.40 20627.80 119.20 10313.90 191.78 6.45 58.92 7.07 76.96
    sdc 0.00 43.40 260.40 112.20 3793.60 29836.40 1896.80 14918.20 90.26 13.53 33.53 2.69 100.08

    Device: rrqm/s wrqm/s r/s w/s rsec/s wsec/s rkB/s wkB/s avgrq-sz avgqu-sz await svctm %util
    sda 0.00 9.20 0.40 0.60 3.20 81.60 1.60 40.80 84.80 0.16 176.80 152.00 15.20
    sdb 0.00 12.40 10.80 28.80 88.00 8328.40 44.00 4164.20 212.54 2.66 67.37 16.16 64.00
    sdc 0.00 17.60 144.80 68.00 1920.00 18190.80 960.00 9095.40 94.51 14.11 70.50 4.70 100.00

  2. #2
    Trusted Penguin Irithori's Avatar
    Join Date
    May 2009
    Location
    Munich
    Posts
    3,439
    Hi and welcome

    The output shows 100% utilization on sdc.
    Which means that this disk is saturated.
    You must always face the curtain with a bow.

  3. #3
    Just Joined!
    Join Date
    Dec 2012
    Posts
    3
    Quote Originally Posted by Irithori View Post
    Hi and welcome

    The output shows 100% utilization on sdc.
    Which means that this disk is saturated.

    That's what I thought it said. I wasn't sure if I was reading it right. Do you know a way to find out what process are killing that drive? Would i get better performance from creating one big raid 5 array rather than two?

  4. $spacer_open
    $spacer_close
  5. #4
    Trusted Penguin Irithori's Avatar
    Join Date
    May 2009
    Location
    Munich
    Posts
    3,439
    On more recent systems with process accounting available, e.g. atop can show a sorted list of processes using the disc. (press d)

    Raid5 is not for performance, it is a datastorage with redundancy. Useable for logfiles, database dumps and the like.
    - If you want performance from hardware side, go for a raid10 with lots of fast spindles (15k disks) along with one or better more capable raidcontrollers (e.g. hp smartarray p8xx) and a few jbods attached.
    - From filesystem side, consider turning off atime as an easy method. There are more options too consider, but this probably is too big for an forum answer.
    - Then analyse the read/write patterns. Is there workload, that can be distributed? e.g. in a database setup: have separate raids for the system (raid1), logs (raid5), dbdumps (raid5), tempdb (raid10), db data (raid10)
    Last edited by Irithori; 12-13-2012 at 05:25 AM. Reason: typo
    You must always face the curtain with a bow.

Posting Permissions

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