home bbs files messages ]

Forums before death by AOL, social media and spammers... "We can't have nice things"

   comp.protocols.tcp-ip      TCP and IP network protocols.      14,669 messages   

[   << oldest   |   < older   |   list   |   newer >   |   newest >>   ]

   Message 12,693 of 14,669   
   Noah Davids to sgifford@suspectclass.com   
   Re: Weird TCP connection establishment b   
   29 Jan 09 19:52:40   
   
   8707725f   
   From: ndav1@cox.net   
      
   sgifford@suspectclass.com wrote:   
   > On Jan 29, 8:09 am, Noah Davids  wrote:   
   >> sgiff...@suspectclass.com wrote:   
   >>> On Jan 28, 9:28 pm, Barry Margolin  wrote:   
   >>>> In article   
   >>>> <26782998-9070-47dd-a23f-52cf3b276...@g39g2000pri.googlegroups.com>,   
   >>>>  "sgiff...@suspectclass.com"  wrote:   
   >>>>> Hello,   
   >>>>> I'm having a weird problem accepting a TCP connection on my Linux Web   
   >>>>> server from a Blackberry client.  Linux server is kernel   
   >>>>> 2.6.18-1.2798.fc6 from Fedora Core 6.  The initial SYN packet from the   
   >>>>> client arrives as I expect.  The server then sends its SYN/ACK and the   
   >>>>> client sends the final ACK in the 3-way handshake, again as I expect.   
   >>>>> But after that, the server keeps retransmitting its SYN/ACK, and the   
   >>>>> client continues to respond with an ACK.  It's as if the server isn't   
   >>>>> accepting the ACK from the client, but I can't figure out why.  After   
   >>>>> awhile the client starts to send data, and is sent a RST in response.   
   >>>>> Here is the packet trace from tcpdump:   
   >>>> This is the behavior you'd see if the server never receives the ACK.   
   >>>> Which machine did you run tcpdump on, the client or the server?  If you   
   >>>> ran it on the client, try again on the server and see whether it   
   >>>> receives the ACK packets.   
   >>> Thanks for the reply Barry.   
   >>> I was running tcpdump on the server, so I can see that the ACK is   
   >>> arriving at the server.  And I'm inferring that the client is   
   >>> receiving the SYN/ACK because it's sending the ACK.   
   >   
   >> First could the sequence number for the client's ACK be incorrect?   
   >> You'll need to dump the entire packet to see it. Second, could iptables   
   >> or some other firewall be strangely configured?   
   >   
   > The TCP sequence numbers are in the packet dump I posted, and to me at   
   > least they look correct.  iptables is installed on the server, but has   
   > no firewall rules.   
   >   
   > Thanks for the ideas, and any others you may have!   
   >   
   > ----Scott.   
   >   
      
   The sequence number of the client's acknowledgment packet is not   
   displayed only the ack number. With no rules I can't see why iptables   
   would be causing a problem but if it is running even without rules it   
   would be worth stopping it just to be sure.   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)   

[   << oldest   |   < older   |   list   |   newer >   |   newest >>   ]


(c) 1994,  bbs@darkrealms.ca