[quagga-dev 3369] Re: [PATCH] non-blocking I/O from client daemons tozebra

Andrew J. Schorr aschorr at telemetry-investments.com
Thu Apr 28 03:21:57 BST 2005


On Fri, Apr 15, 2005 at 12:20:43AM +0100, Simon Talbot wrote:
> Well -- not sure this tells us anything we did not already know, log as
> follows for the testing just undertaken:
> 
> 2005/04/14 23:53:35 BGP: CPU HOG: command took 83730ms: nei
> 134.222.131.125 sh
> 2005/04/14 23:53:35 BGP: CPU HOG: task vty_read (40023f9c) ran for
> 83730ms
> 
> he-gw1-bgpd# sh th cpu
>  Runtime(ms)    Invoked Avg uSecs   Max uSecs  Type Thread
...
>    83740.000         97    863298    83730000 R     vty_read

Could you perhaps try this again with the new code I just checked
into CVS?  It now calculates real (wall-clock) time in addition
to CPU time.  So this might indicate whether the thread is actually
doing computation, or if it is somehow blocking...

Regards,
Andy



More information about the Quagga-dev mailing list