Configuration to avoid dog-pile when cache is empty

Evgeniy Dolzhenko dolzenko at
Mon Jun 7 12:56:34 CEST 2010

Ok, actually I found this behavior surprising since I was thinking
of every request being processed in complete isolation, but here
is the FAQ entry which indirectly documents the feature I believe

On Sat, Jun 5, 2010 at 3:52 PM, Cosimo Streppone <cosimo at> wrote:
> On Sat, 05 Jun 2010 09:18:08 +0200, Per Buer <perbu at>
> wrote:
>> On Fri, Jun 4, 2010 at 2:34 PM, Evgeniy Dolzhenko <dolzenko at>
>> wrote:
>>> Now what I would like ideally to achieve with Varnish is that only
>>> first request hits Application Server and every other request is
>>> "joined" to that request, i.e. will be served as soon as
>>> Application Server responds to the first request and with
>>> the content of the response. I.e. Varnish will have to hold
>>> the connections open and then use response cached by first request
>>> Is that possible? Or is it just a too crazy thing to do?
>> No. It's quite sensible. It's also the default. :-)
> That sounds surprising to me. Can you explain?
> Say that I restart varnish.
> Then 10 clients hit the "/takes_long_time_to_load.html" page at
> the same time.
> What I understand from your reply is that varnish will hold the
> 9 remaining connections until the first one sees the backend
> reply. At that point, the 9 remaining clients will get served
> the cached object.
> Is that really how it works?
> I'm specifically referring to the first time after restart,
> so grace doesn't apply here.
> --
> Cosimo
> _______________________________________________
> varnish-misc mailing list
> varnish-misc at

More information about the varnish-misc mailing list