On 06.Mar.2003 -- 03:45 PM, arturl wrote:
> 
> 
> > > Generally all connections are returned. But increasing number of
> accessing
> > > the pool (by multiplying number of requested xsp pages) results in pool
> > > blocking. Is it a normal situation?
> >
> > Sure, if you have a pool of 10, you may have 10 concurrent
> > users. Every additional user is blocked until a connection is returned
> > to the pool.
> Yes, I understand it. I just wonder, why there's a hangup of the pool  if
> all connections should be returned.

Without knowing the details here, it appears that connections are held
too long for the pool size and incoming rate.

> >OTOH it should only be fractions of a second that a
> > connection is used.
> > Maybe you should just increase the number of
> > connections in your pool.
> I'm trying on that. BTW - is there a formula on counting the optimal size of
> the pool (taking into account number of concurrent connections)?

None that *I* know of. But that doesn't mean much :-)
If your page holds the connection s seconds and you have c incoming
requests per second, you need s*c connections, right? Let's add some
10% slack, that would give ceil(1.1*s*c) Eg. s=0.500 c=5 => 3
But that's probably too obvious to be a good rule ;-)

        Chris.
-- 
C h r i s t i a n       H a u l
[EMAIL PROTECTED]
    fingerprint: 99B0 1D9D 7919 644A 4837  7D73 FEF9 6856 335A 9E08

---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to