FWIW; this seems to be the same issue is under discussion in squid-users
thread "rock store: a bug or ...?". Cc'ing Henrik Lidström and Nikolai
Gorchilov.
Henrik, Nokolai: if you could followup to squid-dev in future about this
please.
On 2014-02-04 05:25, Kinkie wrote:
Hi,
it appears that the large rock merge has introduced a regression,
probably related to the handling of keepalive and/or about finishing
to send an object to a client.
I have confirmed that the issue is NOT present in revno 13200, but it
is present in r13203 (r13201 and r13202 do not compile).
To reproduce, it is enough to use ab on squid, NOT enabling keepalive
(with keepalive enabled, squid works just fine).
$ ab -c10 -n100 -X localhost:3128 http://www.example.com/
Notice: it doesn't make a difference whether a rock cache_dir is used
or whether collapsed_forwarding is enabled or not; the issue is
present regardless.
Confirmed here with AUFS caching and collapsed-forwarding left at
default. The explicit Connection:close request fails to complete and
close when there is only one request on the connection. I have not yet
replicated with multiple requests.
Amos
