Bug 1297378

Summary: The CPU utilization of RHEL6.8 is almost twice as high as that of RHEL5.11 and RHEL7.2 during netperf communication
Product: Red Hat Enterprise Linux 6 Reporter: Guo, Zhiyi <zhguo>
Component: qemu-kvmAssignee: jason wang <jasowang>
Status: CLOSED NOTABUG QA Contact: Virtualization Bugs <virt-bugs>
Severity: medium Docs Contact:
Priority: unspecified    
Version: 6.8CC: ailan, juzhang, michen, mkenneth, qzhang, rbalakri, virt-maint, xfu, zhguo
Target Milestone: rc   
Target Release: ---   
Hardware: x86_64   
OS: Linux   
Whiteboard:
Fixed In Version: Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2016-01-15 07:08:02 UTC Type: Bug
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:

Description Guo, Zhiyi 2016-01-11 10:54:15 UTC 
Description of problem:
The CPU utilization of RHEL6.8 is twice as high as that of RHEL5.11 and RHEL7.2 during communication

Version-Release number of selected component (if applicable):
netperf:ftp://ftp.netperf.org/netperf/netperf-2.7.0.tar.gz
host kernel:2.6.32-598.el6.x86_64
qemu-kvm:qemu-kvm-0.12.1.2-2.483.el6.x86_64
7.2z guest kernel:3.10.0-327.el7.x86_64
6.8 guest kernel:2.6.32-598.el6.x86_64
5.11 guest kernel:2.6.18-398.el5

How reproducible:


Steps to Reproduce:
1.Boot rhel6.8 guest with 17 virtio blk devices and 3 virtio net devices:
/usr/libexec/qemu-kvm -name rhel6.7 -m 4G \
        -smp 4 \
        -cpu SandyBridge,check \
         -vga qxl\
        -spice port=3002,disable-ticketing \
        -device virtio-serial -chardev spicevmc,id=vdagent,debug=0,name=vdagent \
        -serial unix:/tmp/m,server,nowait \
        -device virtserialport,chardev=vdagent,name=com.redhat.spice.0 \
        -drive file=gluster://10.66.10.61/mygluster/30Ggluster511.qcow2,if=none,id=drive-virtio-disk0,format=qcow2,cache=none,werror=stop,rerror=stop,aio=threads,serial=aguoaguoaguo\
        -device virtio-blk-pci,bus=pci.0,drive=drive-virtio-disk0,vectors=0,scsi=off,id=drive-virtio-disk0\
        -monitor stdio \
        -netdev tap,id=idinWyYp,vhost=on -device virtio-net-pci,mac=42:ce:a9:d2:4d:d7,id=idlbq7eA,netdev=idinWyYp \
        -usb -device usb-kbd,id=input0 \
        -netdev tap,id=idinWyYp1 -device virtio-net-pci,mac=43:ce:a9:d2:4d:d7,id=idlbq7eb,netdev=idinWyYp1 \
        -netdev tap,id=idinWyYp2 -device virtio-net-pci,mac=44:ce:a9:d2:4d:d7,id=idlbq7ec,netdev=idinWyYp2 \
        -drive file=gluster://10.66.10.61/mygluster/data1.qcow2,if=none,id=drive-virtio-disk1,format=qcow2,cache=none,werror=stop,rerror=stop,aio=threads\
        -device virtio-blk-pci,drive=drive-virtio-disk1,scsi=off,id=drive-virtio-disk1\
        -drive file=gluster://10.66.10.61/mygluster/data2.qcow2,if=none,id=drive-virtio-disk2,format=qcow2,cache=none,werror=stop,rerror=stop,aio=threads\
        -device virtio-blk-pci,drive=drive-virtio-disk2,scsi=off,id=drive-virtio-disk2\
        -drive file=gluster://10.66.10.61/mygluster/data3.qcow2,if=none,id=drive-virtio-disk3,format=qcow2,cache=none,werror=stop,rerror=stop,aio=threads\
        -device virtio-blk-pci,drive=drive-virtio-disk3,scsi=off,id=drive-virtio-disk3\
        -drive file=gluster://10.66.10.61/mygluster/data4.qcow2,if=none,id=drive-virtio-disk4,format=qcow2,cache=none,werror=stop,rerror=stop,aio=threads\
        -device virtio-blk-pci,drive=drive-virtio-disk4,scsi=off,id=drive-virtio-disk4\
-drive file=gluster://10.66.10.61/mygluster/data5.qcow2,if=none,id=drive-virtio-disk5,format=qcow2,cache=none,werror=stop,rerror=stop,aio=threads\
        -device virtio-blk-pci,drive=drive-virtio-disk5,scsi=off,id=drive-virtio-disk5\
        -drive file=gluster://10.66.10.61/mygluster/data6.qcow2,if=none,id=drive-virtio-disk6,format=qcow2,cache=none,werror=stop,rerror=stop,aio=threads\
        -device virtio-blk-pci,drive=drive-virtio-disk6,scsi=off,id=drive-virtio-disk6\
        -drive file=gluster://10.66.10.61/mygluster/data7.qcow2,if=none,id=drive-virtio-disk7,format=qcow2,cache=none,werror=stop,rerror=stop,aio=threads\
        -device virtio-blk-pci,drive=drive-virtio-disk7,scsi=off,id=drive-virtio-disk7\
        -drive file=gluster://10.66.10.61/mygluster/data8.qcow2,if=none,id=drive-virtio-disk8,format=qcow2,cache=none,werror=stop,rerror=stop,aio=threads\
        -device virtio-blk-pci,drive=drive-virtio-disk8,scsi=off,id=drive-virtio-disk8\
        -drive file=gluster://10.66.10.61/mygluster/data9.qcow2,if=none,id=drive-virtio-disk9,format=qcow2,cache=none,werror=stop,rerror=stop,aio=threads\
        -device virtio-blk-pci,drive=drive-virtio-disk9,scsi=off,id=drive-virtio-disk9\
        -drive file=gluster://10.66.10.61/mygluster/data10.qcow2,if=none,id=drive-virtio-disk10,format=qcow2,cache=none,werror=stop,rerror=stop,aio=threads\
        -device virtio-blk-pci,drive=drive-virtio-disk10,scsi=off,id=drive-virtio-disk10\
        -drive file=gluster://10.66.10.61/mygluster/data11.qcow2,if=none,id=drive-virtio-disk11,format=qcow2,cache=none,werror=stop,rerror=stop,aio=threads\
        -device virtio-blk-pci,drive=drive-virtio-disk11,scsi=off,id=drive-virtio-disk11\
        -drive file=gluster://10.66.10.61/mygluster/data12.qcow2,if=none,id=drive-virtio-disk12,format=qcow2,cache=none,werror=stop,rerror=stop,aio=threads\
        -device virtio-blk-pci,drive=drive-virtio-disk12,scsi=off,id=drive-virtio-disk12\
        -drive file=gluster://10.66.10.61/mygluster/data13.qcow2,if=none,id=drive-virtio-disk13,format=qcow2,cache=none,werror=stop,rerror=stop,aio=threads\
        -device virtio-blk-pci,drive=drive-virtio-disk13,scsi=off,id=drive-virtio-disk13\
-drive file=gluster://10.66.10.61/mygluster/data14.qcow2,if=none,id=drive-virtio-disk14,format=qcow2,cache=none,werror=stop,rerror=stop,aio=threads\
        -device virtio-blk-pci,drive=drive-virtio-disk14,scsi=off,id=drive-virtio-disk14\
        -drive file=gluster://10.66.10.61/mygluster/data15.qcow2,if=none,id=drive-virtio-disk15,format=qcow2,cache=none,werror=stop,rerror=stop,aio=threads\
        -device virtio-blk-pci,drive=drive-virtio-disk15,scsi=off,id=drive-virtio-disk15\
        -drive file=gluster://10.66.10.61/mygluster/data16.qcow2,if=none,id=drive-virtio-disk16,format=qcow2,cache=none,werror=stop,rerror=stop,aio=threads\
        -device virtio-blk-pci,drive=drive-virtio-disk16,scsi=off,id=drive-virtio-disk16\
        -drive file=gluster://10.66.10.61/mygluster/data17.qcow2,if=none,id=drive-virtio-disk17,format=qcow2,cache=none,werror=stop,rerror=stop,aio=threads\
        -device virtio-blk-pci,drive=drive-virtio-disk17,scsi=off,id=drive-virtio-disk17\
2.use taskset to bound vhost to cpu2
# taskset -cp 2 <vhost process ID>
3.use taskset to bound vhost to cpu3
# taskset -cp 3 <qemu process ID>
4.start netsever in guest and netperf client in another host.
5.start sar(in host):
sar -P 2,3 -u 1 100
6.redo step1 ~ step5 with latest RHEL5-11 and latest RHEL 7.2z
7.compare the cpu utilization
Actual results:
The CPU utilization of RHEL6.8 is almost twice as high as that of RHEL5.11 and RHEL7.2 during netperf communication 

Expected results:
The CPU utilization of RHEL6.8 should same as that of RHEL5.11 and RHEL7.2 during netperf communication

Additional info:
netperf statistic:
Rhel 7.2z guest
Average:        CPU     %user     %nice   %system   %iowait    %steal     %idle
Average:          2      0.00      0.00      8.16      0.00      0.00     91.84
Average:          3      0.13      0.00      0.47      0.00      0.00     99.40
Rhel 6.8 guest
Average:        CPU     %user     %nice   %system   %iowait    %steal     %idle
Average:          2      2.71      0.00      1.41      0.00      0.00     95.88
Average:          3      7.60      0.00     20.81      0.00      0.00     71.59
Rhel 5.11 guest
Average:        CPU     %user     %nice   %system   %iowait    %steal     %idle
Average:          2      0.64      0.00      8.23      0.00      0.00     91.13
Average:          3      0.32      0.00      0.90      0.00      0.00     98.77
Comment 2 jason wang 2016-01-15 02:47:33 UTC 
Can you measure the cpu utilization on host?
Comment 3 Guo, Zhiyi 2016-01-15 07:03:27 UTC 
Test on host with kernel:2.6.32-598.el6.x86_64:
Average:        CPU     %user     %nice   %system   %iowait    %steal     %idle
Average:        all      0.03      0.00      0.79      0.04      0.00     99.14
Average:          0      0.20      0.00     10.40      0.00      0.00     89.41
Average:          1      0.00      0.00      0.04      0.01      0.00     99.95
Average:          2      0.00      0.00      0.00      0.00      0.00    100.00
Average:          3      0.00      0.00      0.00      0.00      0.00    100.00
Average:          4      0.00      0.00      0.00      0.00      0.00    100.00
Average:          5      0.00      0.00      0.00      0.00      0.00    100.00
Average:          6      0.03      0.00      0.02      0.00      0.00     99.95
Average:          7      0.00      0.00      0.01      0.50      0.00     99.49
Average:          8      0.01      0.00      0.01      0.00      0.00     99.98
Average:          9      0.07      0.00      0.28      0.00      0.00     99.65
Average:         10      0.01      0.00      0.02      0.00      0.00     99.97
Average:         11      0.00      0.00      0.00      0.00      0.00    100.00

cpu utilization very like result on 7.2z guest and 5.11 guest
Comment 4 jason wang 2016-01-15 07:08:02 UTC 
(In reply to Guo, Zhiyi from comment #3)
> Test on host with kernel:2.6.32-598.el6.x86_64:
> Average:        CPU     %user     %nice   %system   %iowait    %steal    
> %idle
> Average:        all      0.03      0.00      0.79      0.04      0.00    
> 99.14
> Average:          0      0.20      0.00     10.40      0.00      0.00    
> 89.41
> Average:          1      0.00      0.00      0.04      0.01      0.00    
> 99.95
> Average:          2      0.00      0.00      0.00      0.00      0.00   
> 100.00
> Average:          3      0.00      0.00      0.00      0.00      0.00   
> 100.00
> Average:          4      0.00      0.00      0.00      0.00      0.00   
> 100.00
> Average:          5      0.00      0.00      0.00      0.00      0.00   
> 100.00
> Average:          6      0.03      0.00      0.02      0.00      0.00    
> 99.95
> Average:          7      0.00      0.00      0.01      0.50      0.00    
> 99.49
> Average:          8      0.01      0.00      0.01      0.00      0.00    
> 99.98
> Average:          9      0.07      0.00      0.28      0.00      0.00    
> 99.65
> Average:         10      0.01      0.00      0.02      0.00      0.00    
> 99.97
> Average:         11      0.00      0.00      0.00      0.00      0.00   
> 100.00
> 
> cpu utilization very like result on 7.2z guest and 5.11 guest

So it does not looks like an issue for me. For several reasons, measuring the cpu utilization in guest is not accurate. So close this as NOTABUG.

Thanks