ping problem

[shaahin13631363]

hello We have a Nexus device and a Catalyst device that are connected to each other. A server is connected to the Catalyst. From time to time, the ping of the server is interrupted from the outside, and until we pinging IP from the Nexus, the ping of the server is established from the outside. It is not possible. As soon as we ping the IP from the Nexus console, the ping is also established from outside. Can anyone help and guide us?

[balaji.bandi]

how does the the network looks, where is the Server Located, you need to provide lot of information ?

is the server behind NAT, how is your routing arrangement ?

is the ping loss from Internet or Local network ?

do you see any interface errors connected interface switch to switch and switch to server ?

what kind of Server ? (is the Server going sleep mode ?)

[shaahin13631363]

is the server behind NAT?  No 

how is your routing arrangement ? bgp 

is the ping loss from Internet or Local network ?

Ping is interrupted both from the Internet and from the inside. As soon as the ping command is executed from the Nexus switch, the ping is immediately established from the inside and outside.

do you see any interface errors connected interface switch to switch and switch to server ? No

what kind of Server ? (is the Server going sleep mode ?) hpe server esxi installed on 

[balaji.bandi]

Need some more informaton- what nexus model and nexus code running, what Cat switch IOS code running,.

what Esxi version, what is the VM OS Installed ? are you pinging to esxi or inside VM ?

Do you have vswitch or dswitch configured on esxi ?

how about same simulation, if you connected to Cat 2960 switch Pc - is the behavior same ?

test - connect directly Server (test) to switch and see to eliminate the problem.

[shaahin13631363]

nexus : 

Software
BIOS: version 2.6.0
loader: version N/A
kickstart: version 6.0(2)U3(4)
system: version 6.0(2)U3(4)
Power Sequencer Firmware:
Module 1: version v1.0
SFP uC: version 2.12
BIOS compile time: 08/06/2014
kickstart image file is: bootflash:///n3000-uk9-kickstart.6.0.2.U3.4.bin
kickstart compile time: 10/10/2014 10:00:00 [10/10/2014 23:03:20]
system image file is: bootflash:///n3000-uk9.6.0.2.U3.4.bin

************************************************
2960:

Switch Ports Model SW Version SW Image
------ ----- ----- ---------- ----------
* 1 54 WS-C2960X-48TS-L 15.2(7)E11 C2960X-UNIVERSALK9-M

esxi version : 7u3

 what is the VM OS Installed ? linux and windows 

Do you have vswitch or dswitch configured on esxi ? vswitch

are you pinging to esxi or inside VM ? esxi

server directly connected to 2960

[casanavep]

Are the pings directed to the ESXi management IP or the IP of a VM hosted on the ESXi host?
Does the ESXi host and/or the VM have redundant uplinks? If so, have you tried isolating the connection to a single link at a time to see if the issue persists on a specific port or only occurs when multiple ports are active? This could help identify potential problems with load balancing algorithm mismatches or IP to MAC ARP mapping if VM's IP response may switch between layer 3 interfaces and associated MAC addresses.
Finally, are you absolutely sure there's no duplicate IP address assignment on the network?
- Pete

[shaahin13631363]

Are the pings directed to the ESXi management IP or the IP of a VM hosted on the ESXi host? esxi management

Does the ESXi host and/or the VM have redundant uplinks? no

Finally, are you absolutely sure there's no duplicate IP address assignment on the network? yes

[casanavep]

Verify the ESXi host's subnet mask matches that of the network appliances associated VLAN interface serving as the gateway for that subnet. Also verify that the ESXi host's gateway matches that interface IP. I have seen typos and mismatches in either of these settings cause issues like this many times and be missed by several sets of reviewing eyes. Is the ESXi host directly connected to the switch acting as the host's IP gateway? Are there any load balancing or HA algorithms on the VLAN gateway, such as HSRP?

[shaahin13631363]

server connected directly to 2960 and switch 2960 connected to nexus we user bgp via default gateway as routing management all interface on 2960 configure access expect uplink between switches

[MHM Cisco World]

In NSK do you see mac address of server? If you see it are it learn to correct vlan?

İf not 

After NSK ping server are mac learn in correct vlan?

MHM

[shaahin13631363]

Unfortunately, I didn't check the MAC address table before executing the ping command, but after executing the ping command, the Mac, IP and Vlan are displayed correctly.

[MHM Cisco World]

Wait arp aging and check mac address

MHM

[AKK]

It is a NXOS bug. Please upgrade the firmware of switch. We also faced this issue long back.

[jamesbos96602]

make shure all ports are useing swichport  swichport mode access  swichport access vlan < list of vlans> this should solve this probem also do not put any rules in routeing table let them be auto made, makeing them can stop all this and note u do on lot these turn on the router, and pick wich protocall the router will use,  i willing bet anything u trying ping in from ip address u given to port one wont work,   this i found true with most swichs i work with i dont see why yours would be diffrent ,  now how set ip address to swich with out braking it ,   i would set it throw vlan and give it to a port then run cable back to main swich this work for me,