IPFW and FTP client behind NAT

[Julian Elischer]

On 14/2/18 2:35 pm, wishmaster wrote:
> Hi, colleagues.
>
> I have the main server/router and Samba server behind this one. This Samba server at every night sends some data via FTP to another server on the Internet.
> The first remote server is under my power and use about the same configuration as main plus FTPD (port 2112) daemon.
> The second remote server is not in my power and we use is as backup storage and as I know OS is f...ing Linux.
>
> When I connect to the first server and transmit a very big file with transmission duration > 300 sec, the control channel (port pair 36313 <-> 2112) always "recreated" when the expiration timer aim to zero.
>
> root at xxx: ipfw -d show|grep '111.222.230.62'
> 15150       69        5255 (29s) STATE tcp 111.222.230.62 36313 <-> 111.222.13.195 2112 :nts
> 15150   320423   321696704 (300s) STATE tcp 111.222.230.62 60759 <-> 111.222.13.195 49758 :nts
>
> The issue is with the second remote server. When I transmit a very big file, the control channel does not "recreated" and transmitting this file and all the next is always fails.
>
> root at xxx: ipfw -d show|grep '111.222.0.7'
> 03200  2985778  2299927348 (300s) STATE tcp 111.222.0.253 63307 <-> 111.222.0.7 44678 :nts
> 03200       59        4622 (6s) STATE tcp 111.222.0.253 63623 <-> 111.222.0.7 21 :nts
>
> root at xxx: ipfw -d show|grep '111.222.0.7'
> 03200  3137837  2414765852 (300s) STATE tcp 111.222.0.253 63307 <-> 111.222.0.7 44678 :nts
>
> The main server/router uses IPFW and in most places dynamic rules. Is workaround I have added one rule on external interface:
>
> $cmd 5153 allow log tcp from any 21 to any 1024-65535 # ipfw - ftp issue
>
> But I want find the problem.
>
> Thanks,
> Vitaly
> _______________________________________________
> freebsd-ipfw at freebsd.org mailing list
> https://lists.freebsd.org/mailman/listinfo/freebsd-ipfw
> To unsubscribe, send any mail to "freebsd-ipfw-unsubscribe at freebsd.org"
>

can you check the values of the keep-alive timers on all 3 systems?

And possibly the firewall on system3 may block keepalive packets..


[jelischer at bob ~/p4/private/inverness-integ1]$ sysctl 
net.inet.tcp.always_keepalive
net.inet.tcp.always_keepalive: 1

[jelischer at bob ~/p4/private/inverness-integ1]$ sysctl 
net.inet.tcp.keepidle
net.inet.tcp.keepidle: 7200000

that's 2 hours for example.

setting it to less than 300000 should make your control session 
include keepalive packets

also look at your ipfw table and see if this can help you:



      Dynamic rules expire after some time, which depends on the 
status of the
      flow and the setting of some sysctl variables.  See Section SYSCTL
      VARIABLES for more details.  For TCP sessions, dynamic rules can be
      instructed to periodically send keepalive packets to refresh the 
state of
      the rule when it is about to expire.

      See Section EXAMPLES for more examples on how to use dynamic rules.