You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@camel.apache.org by pvenini <pv...@mervaros.com.ar> on 2012/07/02 19:35:17 UTC

High CPU use

Hi, I'm in the process of deploying a Camel solution to it's destination
server. It has 6 routes and is essentially a Fix to HTTP converter (it uses
the FIX and Jetty components). What surprises me is the high CPU use (about
50% continuous on a dual core, dual processor Xeon server) when the router
is idle. 
Is there some config paramether that I should check to see if this could be
improved? I'm using camel 2.7.2 on RHEL 5.

Yours 

Pablo

--
View this message in context: http://camel.465427.n5.nabble.com/High-CPU-use-tp5715368.html
Sent from the Camel - Users mailing list archive at Nabble.com.

Re: High CPU use

Posted by pvenini <pv...@mervaros.com.ar>.
No, the server is not running NTP and has been reset after the leap second.



--
View this message in context: http://camel.465427.n5.nabble.com/High-CPU-use-tp5715368p5721569.html
Sent from the Camel - Users mailing list archive at Nabble.com.

Re: High CPU use

Posted by "Sam (Stephen Samuel)" <sa...@gmail.com>.
This morning we were having the same issue. 150% load on a camel based
routing system that nearly garners only 5% load. I also thought it might
have been time based. The give away for me was seeing a lot of IRQ
interrupts being raised which is often a sign of an out of control timer.

A full restart and back to normal we were. This was a series of AWS servers
btw.

On Mon, Jul 2, 2012 at 10:35 PM, Richard Kettelerij <
richardkettelerij@gmail.com> wrote:

> It's funny you bring this up today. Are you sure it is not the infamous
> Linux leap time bug that's popping up all over the world since last
> weekend? See
> http://destefano.wordpress.com/2012/07/02/java-linux-leap-second-bug/(also
> google 'linux leap time bug').
>
> Essentially it's a bug in the Linux kernel that's producing a CPU load on
> highly concurrent systems. Java based systems often fall in the latter
> category.
>
> Regards,
> Richard
>
> On Mon, Jul 2, 2012 at 7:35 PM, pvenini <pv...@mervaros.com.ar> wrote:
>
> > Hi, I'm in the process of deploying a Camel solution to it's destination
> > server. It has 6 routes and is essentially a Fix to HTTP converter (it
> uses
> > the FIX and Jetty components). What surprises me is the high CPU use
> (about
> > 50% continuous on a dual core, dual processor Xeon server) when the
> router
> > is idle.
> > Is there some config paramether that I should check to see if this could
> be
> > improved? I'm using camel 2.7.2 on RHEL 5.
> >
> > Yours
> >
> > Pablo
> >
> > --
> > View this message in context:
> > http://camel.465427.n5.nabble.com/High-CPU-use-tp5715368.html
> > Sent from the Camel - Users mailing list archive at Nabble.com.
> >
>



-- 
-Sam

Re: High CPU use

Posted by Richard Kettelerij <ri...@gmail.com>.
It's funny you bring this up today. Are you sure it is not the infamous
Linux leap time bug that's popping up all over the world since last
weekend? See
http://destefano.wordpress.com/2012/07/02/java-linux-leap-second-bug/ (also
google 'linux leap time bug').

Essentially it's a bug in the Linux kernel that's producing a CPU load on
highly concurrent systems. Java based systems often fall in the latter
category.

Regards,
Richard

On Mon, Jul 2, 2012 at 7:35 PM, pvenini <pv...@mervaros.com.ar> wrote:

> Hi, I'm in the process of deploying a Camel solution to it's destination
> server. It has 6 routes and is essentially a Fix to HTTP converter (it uses
> the FIX and Jetty components). What surprises me is the high CPU use (about
> 50% continuous on a dual core, dual processor Xeon server) when the router
> is idle.
> Is there some config paramether that I should check to see if this could be
> improved? I'm using camel 2.7.2 on RHEL 5.
>
> Yours
>
> Pablo
>
> --
> View this message in context:
> http://camel.465427.n5.nabble.com/High-CPU-use-tp5715368.html
> Sent from the Camel - Users mailing list archive at Nabble.com.
>