How can I know the reason increase of cpu usage?
Hi ,
I am using RHEL6.5 and am facing cpu usage issue after memory upgrade. I just upgrade physical memory, never changed system. But, my system cpu usage has increased from 11.0%us to 14.6%us. How can I check the reason ?
Before.
top - 11:07:27 up 122 days, 12:23, 1 user, load average: 2.24, 2.28, 2.27
Tasks: 454 total, 2 running, 452 sleeping, 0 stopped, 0 zombie
Cpu(s): 11.0%us, 3.4%sy, 0.0%ni, 85.4%id, 0.1%wa, 0.0%hi, 0.1%si, 0.0%st
Mem: 41273364k total, 40532264k used, 741100k free, 262804k buffers
Swap: 2097144k total, 1582660k used, 514484k free, 7042928k cached
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
29780 root 20 0 26.2g 14g 1.7g S 221.7 36.9 8023:18 nts
3705 root 20 0 7477m 6.6g 357m S 3.2 16.7 250:40.53 qpm
2685 root 20 0 6895m 1.3g 1.3g S 2.5 3.4 2146:50 timestensubd
2671 root 20 0 21.4g 1.3g 1.0g S 1.3 3.4 922:41.07 java
2881 root 20 0 5210m 1.2g 1.1g S 6.9 2.9 4761:53 timestenrepd
7670 root 20 0 4641m 1.0g 1.0g S 0.0 2.6 13:54.27 ttcserver
7090 root 20 0 4814m 126m 121m S 0.0 0.3 15:14.20 subdmodule
5879 root 20 0 672m 64m 62m R 200.3 0.2 352834:51 mrxd
After.
top - 11:16:51 up 5 days, 7:49, 1 user, load average: 2.48, 2.44, 2.44
Tasks: 448 total, 2 running, 446 sleeping, 0 stopped, 0 zombie
Cpu(s): 14.6%us, 4.3%sy, 0.0%ni, 80.9%id, 0.2%wa, 0.0%hi, 0.1%si, 0.0%st
Mem: 66095108k total, 45562896k used, 20532212k free, 333884k buffers
Swap: 2097144k total, 0k used, 2097144k free, 9638268k cached
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
5641 root 20 0 672m 308m 273m R 202.0 0.5 15333:05 mrxd
7237 root 20 0 27.1g 15g 1.8g S 196.1 24.6 11478:19 nts
6990 root 20 0 5212m 1.2g 1.2g S 5.9 1.9 371:18.33 timestenrepd
6539 root 20 0 7500m 7.3g 358m S 3.9 11.6 227:44.85 qpm
1738 root 20 0 0 0 0 S 2.0 0.0 61:31.07 jbd2/sdb1-8
2687 root 20 0 5139m 1.4g 1.4g S 2.0 2.2 160:02.89 timestensubd
5156 root 20 0 27.4g 1.7g 1.0g S 2.0 2.7 80:10.92 java
Thank you.
Responses
One possible explanation is that an application that was previously constrained by memory is now able to execute workloads more efficiently.
Unless you have historical data over a larger period of time, i'm not sure this discrepency is worth the concern.
psacct may be able to give you some insight too (but you really want historical data from before the memory upgrade to compare to)
Guru
6827 points
Bae Daekeuk, of course see the good recommendations from PixelDrift above... If you use psacct, make sure to have some valid logrotate directives; we started using it and noticed it was attempting to fill /var/log and we made appropriate changes to the logrotate directives for psacct logging.
Welcome! Check out the Getting Started with Red Hat page for quick tours and guides for common tasks.