Re: dtrace to trace incoming connection not suceeding ?

Hi!

> >> OK. Can you provide the output of
> >> netstat -sptcp
> >> after some packets were dropped.

> Not sure why you provide two outputs.

I did on the dest host:

netstat -sptcp

then a few

telnet <dest-ip> 179

then a second

netstat -sptcp

That's why I provided two outputs.

There's one small diff between the two that I do not understand:

-       18040 times no signature provided by segment
+       18045 times no signature provided by segment

> Does 'the discarded for bad checksums' counter increase

No 8-(

> If you capture the incoming traffic with Wireshark, does it report that
> the checksum is wrong?

No, it says:

# tcpdump -i vlan500 -n -s 1500 -vvX port 179 and host <source-ip>
tcpdump: listening on vlan500, link-type EN10MB (Ethernet), capture size 1500 bytes
18:24:37.947459 IP (tos 0x10, ttl 64, id 0, offset 0, flags [DF], proto TCP (6), length 60)
    <source-ip>.21692 > <dest-ip>.179: Flags [S], cksum 0x33b8 (correct), seq 2286661000, win 65535, options [mss 1460,nop,wscale 9,sackOK,TS val 286146398 ecr 0], length 0
        0x0000:  4510 003c 0000 4000 4006 0216 d447 c801  E..<..@.@....G..
        0x0010:  d447 c805 54bc 00b3 884b ad88 0000 0000  .G..T....K......
        0x0020:  a002 ffff 33b8 0000 0204 05b4 0103 0309  ....3...........
        0x0030:  0402 080a 110e 3f5e 0000 0000            ......?^....

> > On the same 10g ix0 interface we have three VLANs:
> > - one (vlan551) of them uses tcp-md5 for another bgp session
> > - one (vlan500) does not use tcp-md5, and has the problem
> > - one (vlan724) does not use tcp-md5 for bgp, and works fine
> Possibly the usage of vlans is relevant for this issue. Not sure.

This is the driver we're using:

dev.ix.0.%desc: Intel(R) PRO/10GbE PCI-Express Network Driver, Version - 3.3.24

-- 
pi@FreeBSD.org         +49 171 3101372                  Now what ?