[quagga-dev 6281] Re: BGPD High CP Utilization

Simon Talbot simont at nse.co.uk
Thu Dec 4 17:32:28 GMT 2008

Thanks for that Stephen, the problem is that the routers run on very
small flash drives with a specialised kernel and oprofile is rather
large and can't easily be put on there,


Simon Talbot MEng, ACGI 
(Chief Engineer) 
Tel: 020 3161 6001
Fax: 020 3161 6011

*** Voxige Wholesale Voice http://www.voxige.com *** 

The information contained in this e-mail and any attachments are private

and confidential and may be legally privileged. 

It is intended for the named addressee(s) only. If you are not the 
intended recipient(s), you must not read, copy or use the information 
contained in any way. If you receive this email or any attachments in 
error, please notify us immediately by e-mail and destroy any copy you 
have of it. 

We accept no responsibility for any loss or damages whatsoever arising 
in any way from receipt or use of this e-mail or any attachments. This 
e-mail is not intended to create legally binding commitments on our 
behalf, nor do its comments reflect our corporate views or policies.

-----Original Message-----
From: Stephen Hemminger [mailto:shemminger at vyatta.com] 
Sent: 04 December 2008 16:55
To: Simon Talbot
Cc: quagga-dev at lists.quagga.net
Subject: Re: [quagga-dev 6269] BGPD High CP Utilization

On Thu, 4 Dec 2008 10:13:53 -0000
"Simon Talbot" <simont at nse.co.uk> wrote:

> Hi All,
> I have a very long standing Quagga 0.99.9 router (and yes I know
> upgrading to current is in the plan, but it is a complex border
> For quite a while now, bgpd has been running at near 100% CPU
> utilisation.
> When I look at show work-queues I see something like the following:
>       List (ms)   Q. Runs       Cycle Counts
> P    Items  Hold    Total    Best  Gran.   Avg. Name
>      79020    50 71281551    8376   2794     18 process_main_queue
>          0    50        0       0      1      0 process_rsclient_queue
>          0    10       31       3      3      1 clear
>          0    10       54       5      4      4 clear
>          0    10       27       4      3      3 clear
> .

On Linux with oprofile, you can drill down past that to see where
the hot spot is. On Solaris use dtrace.  (Assuming you haven't stripped
the binaries which is something most distributions do).

Also, look at the network traffic coming in, perhaps there is
something upstream overloading BGP with updates.

More information about the Quagga-dev mailing list