Ethernet Switches & Routers

Reply
N/A
Posts: 1
Registered: ‎01-10-2013

Cannot ping server - ARP Issue

Hi,

I got problem where I cannot ping server from a Cisco switch connected to the FastIron SX 800. It occur in random time..

I have checked the server connection to SX but there is no problem, the server can ping other server and switches in the network.

I attached the detail of topology:

ARP-FastIron800.jpg

SW1 and 2: CISCO WS-C2970G-24TS-E
CORE 1 and 2: Brocade FastIron SX 800, v. 07.3.00bT3e3.

Server: Solaris 10.

I cant ping from switch 2 to server:

SW2#ping 10.43.1.211

Type escape sequence to abort.

Sending 5, 100-byte
ICMP Echos to 10.43.1.211, timeout is 2 seconds:

  .....

Success rate is 0 percent (0/5)

However I can ping from SW1 to server:

SW1#ping 10.43.1.211

Type escape sequence to abort.

Sending 5, 100-byte
ICMP Echos to 10.43.1.211, timeout is 2 seconds:

!!!!!

Success rate is 100
percent (5/5), round-trip min/avg/max = 1/202/1006 ms

The SW2's MAC is not present in the server ARP table.

Server sniffer:

  1. 10.43.1.211 -> (broadcast)  ARP C Who is 10.43.1.35, 10.43.1.35 ?

SW2 sniffer:

Jan  9 09:21:29: IP ARP: rcvd req src 10.43.1.211 0021.281a.6950, dst 10.43.1.35 Vlan100

Jan  9 09:21:29: IP ARP: sent rep src 10.43.1.35 0011.bb91.c141, dst 10.43.1.211 0021.281a.6950 Vlan100

The ARP response never reaches the server.

If I do an clear apr-cache in SW2, then it send a gratiutous ARP message to all network and the ping works.

SW2# clear arp-cache

Jan  9 17:44:20: IP ARP: sent rep src 10.43.1.35 0011.bb91.c141, dst 10.43.1.35 ffff.ffff.ffff Vlan100

Jan  9 17:44:20: IP ARP: rcvd rep src 10.43.1.211 0021.281a.6950, dst 10.43.1.35 Vlan100

Jan  9 17:44:22: IP ARP: creating incomplete entry for IP address: 10.43.1.211 interface Vlan100

Jan  9 17:44:22: IP ARP: sent req src 10.43.1.35 0011.bb91.c141, dst 10.43.1.211 0000.0000.0000 Vlan100

Jan  9 17:44:22: IP ARP: rcvd rep src 10.43.1.211 0021.281a.6950, dst 10.43.1.35 Vlan100

Server sniffer:

10.43.1.35 -> *            ARP C Who is 10.43.1.211, 10.43.1.211 ?

10.43.1.211 -> 10.43.1.35   ARP R 10.43.1.211, 10.43.1.211 is 0:21:28:1a:69:50

10.43.1.35 -> 10.43.1.35   ARP R 10.43.1.35, 10.43.1.35 is 0:11:bb:91:c1:41

Now, the ARP response reaches the server and ping works until the MAC-IP entry expires in server ARP table.

Another workaround:

Adding static entry in server ARP table.

     arp -s 10.43.1.35 00:11:bb:91:c1:41 temp

Its like the DEFECT000357433 bug but its was closed in version 07.3.00:

Defect ID: DEFECT000357433

Technical Severity: High

Summary: Layer 3 traffic sometimes does not get forwarded on an FCX stack.

Symptom: IP Traffic is not forwarded on an FCX stack’s member units.

Workaround: Clearing the MAC or ARP tables temporarily resolves the problem.

Probability: Low

Feature: FCX Layer 3 Forwarding - IPV4

Function: Data Forwarding (IPV4)

Reported In Release: FI 07.2.02

Service Request ID: 601967

Any ideas?

Thanks.

Occasional Contributor
Posts: 5
Registered: ‎04-25-2011

Re: Cannot ping server - ARP Issue

I am seeing a similar issue on code 7.3e on FCXs.  Has any progress been made on this?

Frequent Contributor
Posts: 118
Registered: ‎06-15-2009

Re: Cannot ping server - ARP Issue

sfportero,


The reason could be anything why it’s not working.  It would be best to have a TAC case open so that they can review the configuration, code release, and then try to replicate the issue.  Once replicated, they can isolate the cause and come up with a solution.  It could even be an issue with the cat switches. 

Join the Community

Get quick and easy access to valuable resource designed to help you manage your Brocade Network.

vADC is now Pulse Secure
Download FREE NVMe eBook