unable to ping server

[santhanamanush]

Hello All,

Can I please request your help to resolve the following problem. Here is a quick summary of my lab topology. I have a pair of nexus 5672up switches configured as vpc peers with two 10Gbps link aggregated to form the peer link. The peer keepalive is through direct b2b connectivity between the mgmt.0 ports. Additionally, I have a single 10Gbps link between them setup as a routed port for OSPF adjacency. Lastly I have  a brocade icx switch connected to both these switches using routed ports and with P2P OSPF. all is working except for the following:

I have a physical server which is configured with a vpc (dual homed to both switches and setup with LACP) to both these 5672up switches. the portchannel interface is a trunk and carries multiple vlans. I can ping my storage vlan from both the nexus switches. however I cannot reach it from any of the subnets connected to the brocade switch. the server subnet is visible on the brocade switch and I can ping the HSRP ip of that subnet from the brocade switch. when I do a trace, it reaches the routed port ip on the nexus and drops. so when I did an extended ping to this server ip from the nexus using the routed port ip as the source, it drops.

I am not sure if this is one of those ARP(any kind - proxy, GARP) problems. Any suggestions would be greatly appreciated. I will try to get back with a config but I am looking for leads for me to check before I go with a debug. tks.

[Reza Sharifi]

Hi,

Since you are running OSPF with the brocade switch, can you try adding the server subnet to OSPF as well and test?

HTH

[santhanamanush]

Hi Reza,

Thank you for taking the time to respond. Really appreciated. The server subnet is already being advertised by both the Nexus switches to the Brocade switch. The subnet is very much visible on the Brocade switch and I am able to ping the HSRP virtual IP for the server subnet from the Brocade switch. This HSRP is run between the two Nexus switches. The issue is that once traffic from the Brocade switch reaches the Nexus routed port (directly connected to the Brocade) it appears to go nowhere. I probably need to run debugs. Tks.

[Reza Sharifi]

Hi,

So, from the brocade switch you can ping the VIP for the server vlan, but not the actual servers right?

Are you able to ping the the SVIs on each 5672? Are you running HSRP version 2?

How about ping from the servers to the VIP and SVIs?

Is HSRP working correctly between the switches? "sh hsrp brief" on each switch will tell you what device is active and what device is stand-by.

Can you post configs from the both 5672s?

HTH

[santhanamanush]

Hello Reza, Thank you again. Please refer to the responses below:

brocade switch you can ping the VIP for the server vlan, but not the actual servers right - that is absolutely correct.

SVIs on each 5672? Are you running HSRP version 2 - I can ping both the SVIs from any of the subnets associated with the Brocade switch. I am running HSRP ver2 on both 5672UP switches

How about ping from the servers to the VIP and SVIs - I have asked the relevant teams to test this . This is a virtualized environment and I am sure they will take a bit of time.

HSRP - Hsrp is working as expected. I have explicitly set the priority of one of the switches to 200 and left the other at default(100). the 100 priority switch is in standby mode as expected. looks good on the sh hsrp brief.

--------------------------------

I had a quick two mts to run a couple of debugs including debug ip packet which I tried just as is as well as with a specific destination IP. the interesting thing is that I can see no traffic. I am going to do the same on the brocade tomorrow. really interesting. I am worried if this is either a arp related thing or if one of the vpc related L3 constraints is creeping in. Tks. I will try to get some configs tomorrow.

[jkilleda]

Hello  Anush,

Unable to ping server
Check if the switch MAC is present in the server ARP table