Howdy, Stranger!

It looks like you're new here. If you want to get involved, click one of these buttons!


load average & cpu question
New on LowEndTalk? Please Register and read our Community Rules.

All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.

load average & cpu question

deadbeefdeadbeef Member
edited May 2015 in Help

On an Azure Vm, I see load constantly above 1 (usually 1.5 - 2.5) - 1 core. On htop, the cpu is always low (2-5% most of the time). Tasks 123, 233 thr; 2 running. Memory is half free. OS is ubuntu 14.04

Any idea why load is so high? Could it be a bug? I saw that there used to be an Ubuntu/EC2 kernel bug with false high load, but that was old. On another Azure VM, I don't see the same behavior (with same OS).

Comments

  • joepie91joepie91 Member, Patron Provider
    edited May 2015

    Pending disk I/O, perhaps? Load average takes into account iowait.

    EDIT: Try iotop.

    Thanked by 1deadbeef
  • cociucociu Member

    Overselling ?

    Thanked by 1deadbeef
  • @cociu said:
    Overselling ?

    On Azure? :o

  • deadbeefdeadbeef Member
    edited May 2015

    @joepie91 said:
    Pending disk I/O, perhaps? Load average takes into account iowait.

    EDIT: Try iotop.

    ~30KB/s sporadically atm - load 1.45

  • cociucociu Member

    @deadbeef way do you think is not possible ?

  • @cociu said:
    deadbeef way do you think is not possible ?

    I kinda don't see Microsoft, Amazon and Google having overselling as their profit strategy for their clouds.

  • joepie91joepie91 Member, Patron Provider
    edited May 2015

    deadbeef said: ~30KB/s sporadically atm - load 1.45

    What does it show for % of I/O in use? If the % (ie. random access load) is low, then there's definitely a problem, as the load shouldn't be that high then.

    Does it behave sluggish? I remember Azure having severe disk/network performance issues a year or two ago, when we (as in, ArchiveTeam) tried to use it for archival tasks. I have been told it was improved in the meantime, but who knows.

    Thanked by 1deadbeef
  • deadbeefdeadbeef Member
    edited May 2015

    @joepie91 said:
    What does it show for % of I/O in use? If the % (ie. random access load) is low, then there's definitely a problem, as the load shouldn't be that high then.

    Not sure what the proper params should be, does this include the metric you're referring to?

    $ sudo iotop -bto --iter=5
    08:01:51 Total DISK READ :       0.00 B/s | Total DISK WRITE :       0.00 B/s
    08:01:51 Actual DISK READ:       0.00 B/s | Actual DISK WRITE:       0.00 B/s
        TIME  TID  PRIO  USER     DISK READ  DISK WRITE  SWAPIN      IO    COMMAND
    08:01:52 Total DISK READ :       0.00 B/s | Total DISK WRITE :       0.00 B/s
    08:01:52 Actual DISK READ:       0.00 B/s | Actual DISK WRITE:      19.07 K/s
        TIME  TID  PRIO  USER     DISK READ  DISK WRITE  SWAPIN      IO    COMMAND
    08:01:52   336 be/3 root        0.00 B/s    0.00 B/s  0.00 %  0.14 % [jbd2/sda1-8]
    08:01:53 Total DISK READ :       0.00 B/s | Total DISK WRITE :       0.00 B/s
    08:01:53 Actual DISK READ:       0.00 B/s | Actual DISK WRITE:      15.09 K/s
        TIME  TID  PRIO  USER     DISK READ  DISK WRITE  SWAPIN      IO    COMMAND
    08:01:53   336 be/3 root        0.00 B/s    0.00 B/s  0.00 % 29.41 % [jbd2/sda1-8]
    08:01:54 Total DISK READ :       0.00 B/s | Total DISK WRITE :       3.79 K/s
    08:01:54 Actual DISK READ:       0.00 B/s | Actual DISK WRITE:       0.00 B/s
        TIME  TID  PRIO  USER     DISK READ  DISK WRITE  SWAPIN      IO    COMMAND
    08:01:54 41938 be/3 root        0.00 B/s    3.79 K/s  0.00 %  0.00 % auditd
    08:01:55 Total DISK READ :       0.00 B/s | Total DISK WRITE :       0.00 B/s
    08:01:55 Actual DISK READ:       0.00 B/s | Actual DISK WRITE:       0.00 B/s
        TIME  TID  PRIO  USER     DISK READ  DISK WRITE  SWAPIN      IO    COMMAND
    

    Does it behave sluggish? I remember Azure having severe disk/network performance issues a year or two ago, when we (as in, ArchiveTeam) tried to use it for archival tasks. I have been told it was improved in the meantime, but who knows.

    It only hosts a number of sites (dockerized) that have very very low traffic and some dev sites (which I am not stressing in any way atm). The sites work fine (ok response time on page load).

    I'm thinking of rebooting it to see if that matters but ... but ... the uptime :D

  • joepie91joepie91 Member, Patron Provider

    deadbeef said: Not sure what the proper params should be, does this include the metric you're referring to?

    Doesn't seem to, no.

    If you run iotop without any flags, it should display IO% per process. If all the (top) values are low, there's no notable iowait.

    deadbeef said: It only hosts a number of sites (dockerized) that have very very low traffic and some dev sites (which I am not stressing in any way atm). The sites work fine (ok response time on page load).

    Yeah, then it's probably not an issue with your VM. Perhaps submit a support ticket? No idea how Azure support works...

    Thanked by 1deadbeef
  • vfusevfuse Member, Host Rep

    CPU Steal high?

  • @vfuse said:
    CPU Steal high?

    ?

  • Thank you all for your replies! I'll try to change the node it's on and see if that has anything to do with it.

  • vfusevfuse Member, Host Rep

    @deadbeef said:

    Since it's a VM you could have noisy neighbours in top cpu steal is listed as st.

    Thanked by 1deadbeef
Sign In or Register to comment.