transparent mode -> chksum incorrect

2018-03-22 Thread matei marius
Hello I'm  trying to configure haproxy in transparent mode using the configuration below: The backend servers have as default gateway the haproxy IP (172.17.232.232) frontend fe_frontend_pool_proxy_3128     timeout client 30m     mode tcp     bind 172.17.232.232:3128 transparent    

Re: AW: transparent mode -> chksum incorrect

2018-03-22 Thread matei marius
haproxy -vv HA-Proxy version 1.8.4-1deb90d 2018/02/08 Copyright 2000-2018 Willy Tarreau Build options :   TARGET  = linux26   CPU = generic   CC  = gcc   CFLAGS  = -m64 -march=x86-64 -O2 -g -fno-strict-aliasing -Wdeclaration-after-statement -fwrapv

Re: Can HA-Proxy set an header when he "breaks" stick routing

2018-03-22 Thread Igor Cicimov
Hi, On Thu, Mar 22, 2018 at 6:24 PM, Gisle Grimen wrote: > Hi, > > > > Thank you for your response. > > > > To be very precise the feature I am looking for from HA-Proxy is that when > HA-Proxy does a re-dispatch HA-Proxy also ads a Header, which will tell the > server

AW: transparent mode -> chksum incorrect

2018-03-22 Thread Mathias Weiersmüller
Hi Marius, your NIC is probably doing the TCP checksum calculation (called « TCP offloading»). The TCP/IP stacks therefore sends all outbound TCP packets with the same dummy checksum (in your case: 0x2a21) to the NIC driver. This saves some CPU cycles. Check your TCP offloading settings

Re: transparent mode -> chksum incorrect

2018-03-22 Thread Lukas Tribus
Hello, On 22 March 2018 at 11:49, matei marius wrote: > When I try to access the service from the same IP class with haproxy I see > the packets having incorrect checksum. This is most likely due to offloading techniques such as TX checksumming, where tcpdump will not see

Re: Can HA-Proxy set an header when he "breaks" stick routing

2018-03-22 Thread Igor Cicimov
On Thu, Mar 22, 2018 at 10:42 PM, Igor Cicimov < ig...@encompasscorporation.com> wrote: > Hi, > > On Thu, Mar 22, 2018 at 6:24 PM, Gisle Grimen > wrote: > >> Hi, >> >> >> >> Thank you for your response. >> >> >> >> To be very precise the feature I am looking for from

Re: h2 sending RSTs in response to RSTs

2018-03-22 Thread Willy Tarreau
Hi, On Thu, Jan 25, 2018 at 04:35:58AM +0800, klzgrad wrote: > Hi, > > I patched Chromium to close the stream immediately after seeing END_STREAM. > > In testing, Chromium sends an RST (CANCEL) for this, but HAProxy > replies with an RST (STREAM_CLOSED). This is a MUST NOT (though only a >

Re: AW: transparent mode -> chksum incorrect

2018-03-22 Thread Willy Tarreau
On Thu, Mar 22, 2018 at 01:15:26PM +, matei marius wrote: > haproxy -vv > HA-Proxy version 1.8.4-1deb90d 2018/02/08 > Copyright 2000-2018 Willy Tarreau > > Build options : >   TARGET  = linux26 >   CPU = generic >   CC  = gcc >   CFLAGS  = -m64 -march=x86-64 -O2 -g

Re: Can HA-Proxy set an header when he "breaks" stick routing

2018-03-22 Thread Igor Cicimov
Hi, On Wed, Mar 21, 2018 at 8:57 PM, Gisle Grimen wrote: > Hi, > > Il try to be more specific: > > The functionality I was looking for on HA-Proxy in connection with > sticky-routing is the following: > > Normal flow all servers up (this is functionality available today):

Re: Can HA-Proxy set an header when he "breaks" stick routing

2018-03-22 Thread Gisle Grimen
Hi, Thank you for your response. To be very precise the feature I am looking for from HA-Proxy is that when HA-Proxy does a re-dispatch HA-Proxy also ads a Header, which will tell the server receiving the request from HA-Proxy that HA-Proxy has done a re-dispatch. This is the critical feature

Binding frontends to threads

2018-03-22 Thread Miles Hampson
Haproxy Developers, Is there some way to bind frontends to threads, in the same manner they can be bound to processes via bind-process? I notice the "process" keyword for listener sections accepts a [/], I couldn't find any equivalent in the documention for frontend sections though. Thanks,

Re: actconn issue

2018-03-22 Thread J. Kendzorra
So I tried debugging this a little further; in listener.c I added some debug output around the sections where actconn is in-/decreased. As light test traffic is running, I can see the following (took a few hours): (...) Increased actconn to 1 (in line 611) Decreased actconn to 0 (in line 682)