Dear, Yes, I didn't mention 500 MISS from start, but that the reason I assumed that accept_filter makes different between my old cache and new cache. Now, my old cache doesn't has any 500 MISS but my all new caches have. All cache share the same domain and same original servers. The 500 MISS come from normal accesses as normal link (we're working on reverse proxy scheme) All new caches were rebuilt as you suggested above. Thanks, Trung Kien On Wed, Nov 28, 2012 at 6:54 PM, Amos Jeffries <squid3@xxxxxxxxxxxxx> wrote: > On 28/11/2012 10:37 p.m., Le Trung, Kien wrote: >> >> Hi, >> >> I followed your guide and the WARNING message disappears at start. >> However, new machines still have " 500 " MISS in the access.log while >> old machine do not have any " 500 " MISS. Those caching machines >> (including old and new ones) serve the same domains. >> And those are reverse proxies. > > > "still" ? this is the first time you mentioned any 500 status problem. > > What are the circumstances of the 500 being sent? > and can you repeat the issue to read of the error page Squid sent out with > that status? > > Amos > > >> I'll back tomorrow to inform you if 500 MISS gone. >> >> Thanks and regards, >> Trung Kien. >> >> On Wed, Nov 28, 2012 at 1:36 PM, Amos Jeffries <squid3@xxxxxxxxxxxxx> >> wrote: >>> >>> On 28/11/2012 4:24 p.m., Le Trung, Kien wrote: >>>> >>>> Hi, >>>> >>>> Could anyone get accept_filter working on CentOS 6.3 with squid-3.2.3 ? >>>> >>>> I have already got squid-3.1.18 working properly on CentOS 5.5. >>>> >>>> Start Squid with -d9 I got this: >>>> 2012/11/28 09:19:22 kid1| WARNING: accept_filter not supported on your >>>> OS >>>> >>>> >>> It looks like a amissing include has caused accept_filter to be disabled. >>> Try adding this to src/comm/TcpAcceptor.cc and re-building Squid: >>> >>> #include <netinet/tcp.h> >>> >>> >>> Amos >> >> >> > -- Best Regards, Trung Kiên.
