Home / server / Questions / 655931
Accepted
brendonofficial
Asked: 2015-01-03 08:06:24 +0800 CST

intel xeon e3 1231 v3 server load

  • 772

I'm running a server with a single Intel Xeon E3-1231 v3 processor which is used as a virtual machine host (host OS is CentOS 7 and I'm using the KVM hypervisor via virt-manager).

I created 4 virtual machines, installed CentOS 7, and they're all current running "yum update" simultaneously.

When running "top" or "htop" on the host, I'm getting reported server loads fluctuating between 18.0 and 36.0 roughly, but the host nor any of the running virtual machines show any sign of performance issues at all - in fact everything is still performing great.

Through "htop", I can see all of the 8 cores, and during this time each core is using less than 2% usage.

Either I am completely misunderstanding server load values for an 8-core processor (I'm under the assumption that 8.0 is 100% CPU load), or somehow (if possible) the server load is being read incorrectly?

Is there anything that I can use that will verify that the server loads I am seeing are in fact correct? Could there be some sort of mis-configuration here?

As mentioned, the server is showing no struggles - sensors are showing CPU temperature at an average 32 degrees celsius and a low percentage per core.

Thanks in advance to those who share their thoughts.

Here is the output I am receiving from "top" at the time of edit.

top - 03:09:20 up 1 day, 10:50, 4 users, load average: 20.74, 24.92, 23.20
Tasks: 257 total,   1 running, 256 sleeping, 0 stopped,   0 zombie
%Cpu(s):  0.6 us,  0.2 sy,  0.0 ni, 93.2 id,  6.1 wa,  0.0 hi,  0.0 si,  0.0 st
KiB Mem:  16211176 total, 10311624 used,  5899552 free,      804 buffers
KiB Swap: 10239996 total,        0 used, 10239996 free.  1233840 cached Mem
centos

1 Answers

  1. Best Answer

    I had eventually worked out that the displayed load was being caused by the speed of the hard drives, and the RAID set up I was using. I had been using BIOS-based RAID (fakeRAID) with two 7,200RPM Toshiba drives, and it appears as though the drives were not being written to fast enough, causing CPU back log.

    Just using iotop on it's own didn't show any stress on the hard drives, but rather I had to utilise the iostat tool which was showing the drives at ~98.0-100.0 load.

    Since purchasing a proper Adaptec RAID card, and two 15K RPM drives the load issue has since not occurred.

    • 2