Forum Discussion

schulerb_82266's avatar
schulerb_82266
Icon for Nimbostratus rankNimbostratus
Feb 02, 2008

Forwarding Virtual Server not working

I am attempting to have hosts on the internal vlan connect to hosts on the other side of the F5.

 

 

------

 

65.x.x.x untrust

 

11.1.1.1 trust

 

 

11.1.1.20 self-ip of external

 

10.1.1.20 self-ip of internal

 

 

10.1.1.130 server

 

------

 

 

The BigIp has a default gateway to 11.1.1.1. The web-server (10.1.1.130) has the gateway as 10.1.1.20.

 

I have a vip on the F5 (10.1.1.25) that hits 10.1.1.130:80.

 

 

Everything seems to work ok. A host at 11.1.1.99 can hit 11.1.1.20 and 10.1.1.25 ok.

 

 

The F5 can see 11.1.1.1 and beyond (internet) and 10.1.1.130 and the vip 10.1.1.25.

 

 

The web-server can get to 10.1.1.20, but not beyond.

 

 

I have created a forwarding virtual server on the F5.

 

 

virtual forward-all {

 

ip forward

 

destination any:any

 

mask none

 

}

 

 

But, no luck. After looking at a load of these posts, I think I am on the right track, but am missing something obvious.

 

 

Any gotchas that I am missing?

 

 

Ben
  • Hi Ben,

     

     

    can you try to create a gateway_pool with F5 gateway as the pool member, associate this pool to a wildcard_VS. the wildcard_VS should use Performance Layer 4 and select All protocols.

     

     

    pool gw_pool {

     

    members 11.1.1.1:any

     

    }

     

    virtual wildcard_vs {

     

    pool gw_pool

     

    destination any:any

     

    mask none

     

    }

     

     

    if still not working, you can try to use tcpdump to capture both vlan traffic.

     

     

    regards,

     

    KY

     

  • Thanks for the reply KYY. I've implemented your suggestion. Unfortunately, I am in the same spot.

     

     

    From the bigip.conf:

     

     

    route default inet {

     

    gateway 11.1.1.1

     

    }

     

    pool gw_pool {

     

    member 11.1.1.1:any

     

    }

     

    virtual ip_forward {

     

    destination any:any

     

    pool gw_pool

     

    mask none

     

    }

     

     

    On the F5, I ran 'tcpdump -n -i 0.0 tcp port 8093'.

     

    From 10.1.1.130 with its gateway of 10.1.1.20, I ran to a known working IP -- 'telnet 12.162.xx.xxxx 8093'.

     

     

    The F5 captured:

     

     

    17:49:19.399399 802.1Q vlan4094 P0 10.1.1.130.36542 > 12.162.xx.xxxx.8093: S 168

     

    875122:168875122(0) win 5840 2> (DF) [tos 0x10]

     

    17:49:19.399468 802.1Q vlan4093 P0 10.1.1.130.36542 > 12.162.xx.xxxx.8093: S 168

     

    875122:168875122(0) win 5840 2> (DF)

     

    17:49:22.399520 802.1Q vlan4094 P0 10.1.1.130.36542 > 12.162.xx.xxxx.8093: S 168

     

    875122:168875122(0) win 5840 2> (DF) [tos 0x10]

     

    17:49:22.399546 802.1Q vlan4093 P0 10.1.1.130.36542 > 12.162.xx.xxxx.8093: S 168

     

    875122:168875122(0) win 5840 2> (DF)

     

    17:49:27.062530 802.1Q vlan4094 P0 12.162.xx.xxxx.8093 > 10.1.1.130.36542: R 0:0

     

    (0) ack 168875123 win 0 (DF)

     

     

    4093 is the external vlan and 4094 is the internal.

     

     

    If I am reading this correctly, the .130 hosts connects to the internal vlan, which connects to the external, but then nothing more.

     

     

    From the F5, I am able to run the telnet command successfully. I don't think it is routing, but I am in need of a clue as to what to look at next.

     

     

    This is the same results I got with my original configuration attempt.

     

     

    Ben
  •  

    I think I have a switch issue (between the web-server and the F5). On the web-server, I explicitly set the interface to 100/full/autoneg off and things started working.

     

     

    I don't get it.

     

     

    Ben