Very obscure one way routing issue 5406ZL

[ipconflict]

Dear all.

5406ZL

Image stamp:    /sw/code/build/btm(kickers_t4)
                Jul  7 2009 13:00:45
                K.14.34
                189
Boot Image:     Primary

We have a very very strange issue.

We have a citrix farm with 20 Virtual Windows terminal servers, all on DHCP, All get an IP in the range 128.129.12. x as a resevervation. All the same gateway and subnet.  All get the same thinly Provided boot IMAGE and boot cleanly every morning.

The citrix farm is on a 4 port bond into our 5406, which also has our proxy server into it.

All bar one of the servers can ping our proxyserver.

That single one can ping everything except the proxy server.

The proxy server can ping that one server ( and all other devices can ping the server)

The proxy can't traceroute to the server.  "Network Down"

The one server gets a trace to the (5406) and then fails to go to the Proxy server.

The proxy vendor ( Sophos ) assure me that they don't do any firewalling.

Wireshark captures on the server show the ICMP request but no response.

I am writing this in the hope that someone may have a glimer of an idea, the only thign other than restarting the switch that I have not done is restart the proxy ( the one that I am assured is not a problem by the vendor)

Rob

[Apachez-]

I guess you have checked the obvious?

ip/mask and gateway of all devices?

Routingtable of all devices?

arp table of all devices (do you see expected values)?

[Chrisd131313]

Hi ipconflict,

Have you tried clearing the mac address for teh one server which seems to have teh issue from the switches mac-address table? 

You will have to clear down the whole VLAN or port table to do this, but I have seen issues on the 5406zl similar to what you describe. 

clear mac-address vlan <vlan_num>

clear mac-address port <port_num>

HTH

[ipconflict]

After scratching my head, checking, checking and checking everyhting up down left right and in every other imaginable dimsion I came to the conclusion that I had to try a restart on the proxy device.

Prior to this I had tried a ping from the proxy to the server which worked fine.

And also a traceroute which resulted in the proxy reporting "Network Down"

I rebooted the proxy device and normal service was restored.

The prox vendor still maintains that the device has no blocking of ICMP/TCP etc.  But I can't come to any other logical conclusion.

Rob