[ULTRAMONKEY-USERS] delayed activation of backend server

[Simon Horman]

On Thu, Jun 19, 2008 at 03:13:22PM +0200, Wichert Akkerman wrote:
> For Zope clusters it would be very practical to be have backendservers 
> stay out of the active pool until they have been active for a period of 
> time. Similar to how it checkcount determines how often a service must 
> have failed its check before it is considered dead it would be nice ot 
> have a count that determines how often a check much succeed before a 
> server is considered to be up again. The use case for us is that we want 
> some extra time to warm up caches in a (re)started Zope server before 
> adding it to the pool.
> 
> Is this a planned feature for UM?

Hi Wichert,

that sounds fairly straightforward. I hadn't planed such a feather,
but it should be easy enough to add. But I am curious, why is
such a thing desirable for Zope clusters?

-- 
Horms