Re: [jetty-users] [External] Re: jetty hot/stuck thread

[Lu Tahmazyan <ltahmazyan@xxxxxxxxxxxxx>]

The reason why I think this is a stuck thread is this.  There are 22 instances of the application running all except one showing similar CPU usage, the one with stuck thread is running ~50% hotter (see image).  Note that it runs hot for days until that offending instance is bounced.

Looking at threads via VisualVM we see stuck thread is constantly running 

We are currently testing/validating Java 11.0.3 for production use, will move to it shortly after which Jetty 9.4.18 will be tested next. 

Thanks, 

-Lu 

In case images don't come through: 

cpu: https://opentable.d.pr/e8oBF9

thread: https://opentable.d.pr/xk9FMc

 

On Fri, May 3, 2019 at 10:06 AM Simone Bordet <sbordet@xxxxxxxxxxx> wrote:

Hi,

On Wed, May 1, 2019 at 7:31 PM Lu Tahmazyan <ltahmazyan@xxxxxxxxxxxxx> wrote:
>
> Hello there.
>
> We have been running a reverse proxy based on Jetty's AsyncMiddleManServlet for years without any major issue.  However, we started noticing that once in a great while a thread gets stuck in a hot loop.  The only way to recover it is to bounce the application.
>
> "Hot" part of the code seems to be around these methods calls (this was verified via JMX CPU sampling)

The stack traces you show seem completely normal to me.
How do you know it was stuck in a loop?

Would be great if you can update to JDK 11 and Jetty 9.4.18.

--
Simone Bordet
----
http://cometd.org
http://webtide.com
Developer advice, training, services and support
from the Jetty & CometD experts.
_______________________________________________
jetty-users mailing list
jetty-users@xxxxxxxxxxx
To change your delivery options, retrieve your password, or unsubscribe from this list, visit
https://www.eclipse.org/mailman/listinfo/jetty-users