2960G's not communicating

Answered Question
Jan 6th, 2009
User Badges:
  • Bronze, 100 points or more

We have to 2960g'S and from switch 0 I can't ping switch 1 and people on switch 1 can't access the proxy server on switch 0


I can't see anything wrong with this config please help


and as there was only one cable going between the old switches there is only one going between the new switches and it happens to be going from sw0/47 to sw1/48



Correct Answer by sachinraja about 8 years 5 months ago

Richard. Thats cool. But any idea why the portchannel did not work ? Layer 2 on portchannel was fine, but no layer 3 access ? were you able to troubleshoot that ?


Anyway, great to see that you have solved the issue...


Raj

  • 1
  • 2
  • 3
  • 4
  • 5
Overall Rating: 4.8 (10 ratings)
Loading.
switchtower Tue, 01/06/2009 - 09:25
User Badges:

From what I can see, the problem you might have is with how the port-channel is setup. Currently, on both switches, you are using ports Gi0/47 and 48 in the port-channel.


If you don't plan on using port channel, I would remove that portion of the configuration and just have normal trunk connections between sw0/47 to sw1/48.


You can remove the port-channels by doing the following on both switches:


interface range Gigabitethernet 0/47-48

no channel-group 1 mode on

no interface port-channel 1


The interfaces will be just normal trunk ports now, since the command "switchport mode trunk" was already there.



Jon Marshall Tue, 01/06/2009 - 09:27
User Badges:
  • Super Blue, 32500 points or more
  • Hall of Fame,

    Founding Member

  • Cisco Designated VIP,

    2017 LAN, WAN

Please post output of


1) sh int trunk

2) sh ip int brief

3) sh cdp neigh


from both switches.


Jon

is66rlhntadm Tue, 01/06/2009 - 09:50
User Badges:
  • Bronze, 100 points or more

As the switches are over 4,000 miles away from me and there isn't a console cable going to SW1 I can only sh the output from SW0

(I will have access to SW1 tommorow but I need to figure out what part of the config I messed up on)


if simpily removing the port-channel would work I'll remove it from both but we want the port-channel for speed and reliability(although the 2nd cable wasn't connected yet)



Attachment: 
is66rlhntadm Tue, 01/06/2009 - 10:23
User Badges:
  • Bronze, 100 points or more

And I still can't ping 172.18.18.3 from ether the server or sw0

sachinraja Tue, 01/06/2009 - 10:35
User Badges:
  • Red, 2250 points or more

Hey richard


Are you getting the mac-address/ARP details of 18.3 from Switch 0 ?? Is the layer 2 communication through ? Might be there is another PC or device with the same IP address, and there could be a IP clash ! just guessing.. Troubleshoot starting from layer 1, then 2, and atlast 3 ! Are there any logs, in "show log" output ??


Also, add the command "switchport encapsulation dot1q" on the trunk interfaces, to force dot1q.. In anycase, your native vlan is 1, and the packets would go untagged, just like an access port....


Hope this helps..


Raj

is66rlhntadm Tue, 01/06/2009 - 10:49
User Badges:
  • Bronze, 100 points or more

the 2960G's don't have the command "switchport encapsulation dot1q" (as they can only do Dot1Q)


and in the output of show Mac address-table shows multiple mac addresses from interface PO1



sachinraja Tue, 01/06/2009 - 10:52
User Badges:
  • Red, 2250 points or more

does it show the mac-address of 18.3 from Switch 0 , in the ARP table? If yes, can you see if that particular ARP is forwarded out of PO 1, using the show mac-address table command ?


show arp for 18.3, take the ARP and give a show mac-addr table..


Raj

is66rlhntadm Tue, 01/06/2009 - 10:56
User Badges:
  • Bronze, 100 points or more

I don't know off hand a mac address for SW1(18.3) and till tommorow I can't connect to the Command line


as for IP addresses We've used the same ones that the former switches used (which were swaped out) so there shouldn't be an IP conflict


users on sw1 can't reach users on sw0

is66rlhntadm Tue, 01/06/2009 - 10:59
User Badges:
  • Bronze, 100 points or more

After pinging (unsuccessfully of course) 18.3 from SW0 then quickly doing a SH arp it shows incomplete as the mac address for 18.3

sachinraja Tue, 01/06/2009 - 11:08
User Badges:
  • Red, 2250 points or more

Can you configure switchport mode access, and switchport access vlan 1 on the gig0/47 port of switch 0 connecting to Switch 1, for troubleshooting purpose , and remove trunk ? Let it act as a normal access port and try to reach VLAN 1's ip address on the second switch !


in anycase, we will have to wait till u get access to second switch !


Raj

sachinraja Tue, 01/06/2009 - 11:01
User Badges:
  • Red, 2250 points or more

No.. what I was asking you to check is, if you have an ARP table entry for 172.18.18.3 on Switch 0.. If this doesnt exist, it is a layer 2 issue... You can probably do a clear ARP on switch 0, to make sure the IP 18.3 doesnt point to the older switch's MAC !! ARP will re-build again, and populate the mac address table !


If you dont have any ARP for 18.3 on Switch 0 (using the show arp command), it is tough to troubleshoot, unless you get access to second switch...


Raj

is66rlhntadm Tue, 01/06/2009 - 11:08
User Badges:
  • Bronze, 100 points or more

There is no entry on SW0 for 18.3 in the arp table


but if I try to ping 18.3 an entry shows up for a few seconds (but is listed as incomplete)

glen.grant Tue, 01/06/2009 - 11:00
User Badges:
  • Purple, 4500 points or more

If he removes the portchannel then one link will become blocked because that creates a built in loop so it will be a single link carrying traffic between them . Is layer 1 ok ? Verify with "show interface status" , 47 and 48 should show connected. Does a show interface trunk look right , though this can be deceiving when the trunks and channels are forced on . That being said even if the trunk was working who is doing the layer 3 routing as the 2960's are layer 2 only and need a layer 3 device to route between the different vlans.

sachinraja Tue, 01/06/2009 - 11:03
User Badges:
  • Red, 2250 points or more

glen.. layer 1 seems to be fine.. had a look at CDP !


SW0#sh cdp n

Capability Codes: R - Router, T - Trans Bridge, B - Source Route Bridge

S - Switch, H - Host, I - IGMP, r - Repeater, P - Phone


Device ID Local Intrfce Holdtme Capability Platform Port ID

SW1 Gig 0/47 150 S I WS-C2960G Gig 0/48


it does show the second switch connected.. cdp is fine.. need access to second switch i guess..


Raj

is66rlhntadm Tue, 01/06/2009 - 11:06
User Badges:
  • Bronze, 100 points or more

The installer only put in one cable (goes sw0/47-to-sw1/48)


as I'm quite far away from the switches I can't easily connect another cable


yes sh int trunk looks ok


would "channel-group1 mode desirable" be a more robust config


we would like to use etherchannel but can get by without

glen.grant Tue, 01/06/2009 - 11:09
User Badges:
  • Purple, 4500 points or more

A negotiated port channel does check the port paramters and will only bring the channel active if everything is configured ok . Everyone has there own opinion on whether to force on the channels and trunks . We usually use pagp or lacp to create the etherchannels.

sachinraja Tue, 01/06/2009 - 11:11
User Badges:
  • Red, 2250 points or more

Can you configure switchport mode access, and switchport access vlan 1 on the gig0/47 port of switch 0 connecting to Switch 1, for troubleshooting purpose , and remove trunk ? Let it act as a normal access port and try to reach VLAN 1's ip address on the second switch !


in anycase, we will have to wait till u get access to second switch !


Raj

is66rlhntadm Wed, 01/07/2009 - 07:59
User Badges:
  • Bronze, 100 points or more

I ended up just removing the port-channel and doing a normal trunk

Correct Answer
sachinraja Wed, 01/07/2009 - 08:02
User Badges:
  • Red, 2250 points or more

Richard. Thats cool. But any idea why the portchannel did not work ? Layer 2 on portchannel was fine, but no layer 3 access ? were you able to troubleshoot that ?


Anyway, great to see that you have solved the issue...


Raj

is66rlhntadm Wed, 01/07/2009 - 08:11
User Badges:
  • Bronze, 100 points or more

It looks to have been a layer 2 issue


the arp table would only ever show incomplete for 18.3 so I think the "ON" mode isn't as flexable I thought


maybe if there had been a 2nd cable

if/when the ship gets back up here I'll see about applying an ether channel again

Actions

This Discussion