Link to home
Start Free TrialLog in
Avatar of svasilakos
svasilakos

asked on

SonicWall issues a TCP Reset periodically on an open connection

We have several server based applications that open a TCP connection to remote servers for messaging back and forth.  Some of the remote hosts we are connecting too are accessed via routers on the same lan segment and some are accessed via routers on a different lan segment off of the same Pro4060.

What we are experiencing is if we use the SonicWall as the default gateway to forward packets to the routers that provide access to the hosts, the SonicWall will reset the TCP connection exactly in 15 minute intervals. If we put static routes on our servers to forward packets directly to the routers ( by passing the sonicwall) than the connection remains constant. We can obviously put static routes on the serves where the routers are on the same segments but when the routers are on different LAN segments we have to go through our 4060

The TCP Connection timeout on the Pro4060 is set to 60 minutes. We have not really played with other settings. We are just wondering if there is some configuration setting we are missing or why SonicWall is RST'ing the connection. We have also experienced this with certain applications in the middle of a message flow.

Running wireshark on the servers we see the following packet after 15 minutes.

No.     Time        Source                Destination           Protocol Info
      8 912.207429  172.16.13.140         180.20.12.16          TCP      48004 > 4286 [RST] Seq=1 Win=16560 Len=0

Frame 8 (62 bytes on wire, 62 bytes captured)
Ethernet II, Src: Sonicwal_02:cf:42 (00:06:b1:02:cf:42), Dst: Vmware_d9:f0:b2 (00:0c:29:d9:f0:b2)
Internet Protocol, Src: 172.16.13.140 (172.16.13.140), Dst: 180.20.12.16 (180.20.12.16)
Transmission Control Protocol, Src Port: 48004 (48004), Dst Port: 4286 (4286), Seq: 1, Len: 0
    Source port: 48004 (48004)
    Destination port: 4286 (4286)
    [Stream index: 1]
    Sequence number: 1    (relative sequence number)
    Acknowledgement number: Broken TCP. The acknowledge field is nonzero while the ACK flag is not set
    Header length: 20 bytes
    Flags: 0x04 (RST)
    Window size: 16560
    Checksum: 0x3948 [validation disabled]
    [SEQ/ACK analysis]

Does anyone know what this might be? I can provide the whole packet capture if needed.

Thanks












Avatar of digitap
digitap
Flag of United States of America image

So, if you have different segments, then you must have firewall rules for those segments.  If so, check the TCP timeouts configured for the firewall rules under the advanced tab.
Avatar of svasilakos
svasilakos

ASKER

Well what do you know. The LAN > LAN has a TCP connection reset of 15 minutes. Thanks, now one question.  This is a rule auto added for any to any traffic and the TCP connection timeout is grayed out so I cannot change? Do I need to setup a special rule for certain traffic or is there a way I can change the Any to Any...

-Steve
ASKER CERTIFIED SOLUTION
Avatar of digitap
digitap
Flag of United States of America image

Link to home
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
Start Free Trial
Have not verified this solution as it will take some time to change production firewalls. But it looks like that is my problem. Thank you
No problem.  You should post back when you have confirmation.  Thanks for the points!