troubleshooting network problem

[Mohd Khairul Nizam]

In my organization we have several 2950T that conected each other. We have 4 floor with each floor have 1 sw except 4th flr have 3 sw.

Main server, apps, db ( 15 units) connected to core sw at 4th flr.

Currently the problem is others devices can't connected to the server at core sw.

Test ping and the network to core sw is intermittent problem. When ping 100 times only 30 reply. After 10 minutes the connection restored. Ping 100 times successful. It will not having problem again until the next day. This already happen 3 days.

Any command for me to use to check if port or the switch having problem? Any log i can collect

[Leo Laohoo]

Errrrr ... your "core" switch is a 2950? 

What kind of servers do you have?  Can you provide the model numbers?

[Mohd Khairul Nizam]

Hi,

Actually the core is 2970 but i swap with 2950 3 days ago. I thought the core is having problem. So now is 2950, seem like the 2970 is not having problem because the problem still occurs.

conecction to core :- router, asa5510, NEC express 5800/110Gd dserver (db, apps)

[John Blakley]

To me, it sounds like you have a loop somewhere in your network. Personally what I would do is put a host on the switch that the server is connected to and disconnect any uplinks to this switch to leave only this switch, server, and host. Test your ping. If it works fine, then uplink one switch and do the ping again. If it still works, repeat the process until you find the switch that's causing the issue. Once you find that switch, go to it and put your host on this switch and remove all other connections except the uplink. If you're still clean, start reconnecting hosts one at a time until you find the one causing the loop. Then track it down to probably see that someone looped a small switch at their desk

HTH,
John

*** Please rate all useful posts ***

[Mohd Khairul Nizam]

Hi John Blakey,

When u mention 'uplink' did u mean trunk? backbone?

TQ

[John Blakley]

Yes, any links that go to your core from other switches. Once you remove those, you've isolated the core switch to a single switch. If you don't notice the problem there, start reconnecting each other switch one by one.

HTH,
John

*** Please rate all useful posts ***

[paul driver]

Hi Mohd,

If its not feasible to unplug all the uplinks - you can perform roughly the same procedure as John highlighted - apply a continuous ping to your server from a host and then from the core , unplugged each uplink separately checking your ping responses until they are replying fine.

Then whatever link you had unplugged investigate further down the connection until you fine the culprit causing the loop.

After that i would suggest applying some  loop prevention.

res

Paul

[Mohd Khairul Nizam]

Hi all,

After troubleshooting, the suspect is now the core sw itself.

During around 4 ~ 5 pm, everybody cant access to the devices that connect to the core sw.

From remote site, the ping only success until router, the it will fail on the core sw. After 10 minutes the core sw can be ping normally.

For reference, we just plug in 1 server to this core sw last weekend. This intermittent only happen just when this new server installed.

The server is connected using intFa 0/18 switchport mode access (vlan 500).

I want to monitor the traffic of this port. Is there any commnad to use?? I want to know the traffic