Hi all,
I am trying to use xenmon.py to monitor cpu utilization of my machine.
Machine has 2 cores.
[root@xxx script]# xm vcpu-list
Name ID VCPU CPU State Time(s) CPU Affinity
Domain-0 0 0 0 -b- 73753.3 any cpu
Domain-0 0 1 1 r-- 63252.4 any cpu
webserver 4 0 1 -b- 5761.9 any cpu
I ran xenmon.py with interval time set to 5 seconds and and let it run
indefinitely and log into local files by executing the command:
[root@xxxx script]# xenmon.py -n -i 5000 -t 0
The log files xenmon automatically generates are "log-dom0.log"
"log-dom2.log" "log-idle.log", then the result will be in
log-dom2.log looks like the following:
[root@xxxx script]# more log-dom2.log |awk ''{print $1, $2, $5,$9,
$12}''
#passed cpu cpu(%) blocked(%) waited(%)
0.000 0 0.000 0.000 0.000
11.470 1 0.000 0.000 0.000
5013.090 1 1.919 98.073 0.006
5015.670 0 0.104 52.173 0.092
10014.038 1 1.957 45.017 0.055
10016.627 0 1.473 75.309 0.004
15014.840 1 0.460 21.995 0.001
15017.386 0 0.782 37.692 0.116
20015.507 1 1.303 59.982 0.003
20018.086 0 0.208 15.147 0.003
25016.194 1 1.722 83.039 0.005
25018.832 0 0.664 78.986 0.151
30017.152 1 1.467 20.051 0.085
It has so many cases that cpu(%)+blocked(%)+waited(%) not equal or
even close to 100%. I read "XenMon: QoS Monitoring and Performance
Profiling Tool" paper, but still can not figure it out. Especially,
what does the waited(%) mean?
I even find waited(%) more than 100 in some case. How to explain all these?
cpu(%) blocked(%) waited(%)
25.316506 67.791463 176.436726
Thank you in advance
Best,
Lei
_______________________________________________
Xen-users mailing list
Xen-users@lists.xensource.com
http://lists.xensource.com/xen-users