Re: [jetty-users] CPU spikes caused by Selector threads

[Greg Wilkins <gregw@xxxxxxxxxxx>]

There have been jetty bugs with spinning selectors and one was fixed in 8.1.17.  So you could try upgrading to  jetty-8.1.19.v20160209 at least.

On 31 May 2016 at 11:50, Kohei Nozaki <kohei.nozaki@xxxxxxxxxx> wrote:

Hi Greg, thank you so much for your help.

On 5/30/16 8:33 PM, Greg Wilkins wrote:
> A 100% CPU spike might suggest a bug, but a 65% spike could just be busy
> IO or even a DOS attack.

I have observed that there are some samples which indicates 100% CPU (i.e. 0.0%id) during the spike as follows:

$ grep 0.0%id top_h.txt
Cpu(s): 21.3%us, 37.1%sy,  0.0%ni,  0.0%id,  0.0%wa,  0.0%hi,  0.5%si, 41.1%st
Cpu(s): 23.8%us, 34.0%sy,  0.0%ni,  0.0%id,  0.0%wa,  0.0%hi,  0.5%si, 41.7%st
Cpu(s): 25.1%us, 32.9%sy,  0.0%ni,  0.0%id,  0.0%wa,  0.0%hi,  0.5%si, 41.4%st
Cpu(s): 24.4%us, 33.9%sy,  0.0%ni,  0.0%id,  0.0%wa,  0.0%hi,  0.3%si, 41.5%st
Cpu(s): 21.4%us, 36.5%sy,  0.0%ni,  0.0%id,  0.0%wa,  0.0%hi,  0.8%si, 41.4%st

Do these samples might suggest a bug? do you think of any relevant bugs that have been fixed in the latest Jetty?

Thanks,
Kohei
_______________________________________________
jetty-users mailing list
jetty-users@xxxxxxxxxxx
To change your delivery options, retrieve your password, or unsubscribe from this list, visit
https://dev.eclipse.org/mailman/listinfo/jetty-users

--

Greg Wilkins <gregw@xxxxxxxxxxx> CTO http://webtide.com