I think I found the TCP_SWAP_FAIL_MISSS cause.
I was playing with store logs and I have found in section 79 some debug
saying something about the inconsistency of the URL stored in the cached
object metadata at StoreMetaURL.cc.
what I will do is like in the client_side_reply.cc change the
check\disable it in cases that a store_url was done on the request.
In a case this is the last blockage in this area it seems to me kind of
realistic for store_url to be in one of next 3.3 beta versions.
Eliezer
On 10/24/2012 1:50 AM, Eliezer Croitoru wrote:
This is not fully cleanedup patch but it did everything before the update.
no I did some things such as resolving things to make sure it will run
on trunk.
everything now compiles but I have one annoying problem:
in memeory the cache works fine and the problems comes in a form of
TCP_SWAP_FAIL_MISSS and causing the request to be fetched from source.
the main thing in this diff is to find the main reason that can cause
the TCP_SWAP_FAIL_MISS
Regards,
Eliezer
--
Eliezer Croitoru
https://www1.ngtech.co.il
IT consulting for Nonprofit organizations
eliezer <at> ngtech.co.il
