high latency on switch

[arvindai]

In my organisation setup I have 7-8 switches cascaded each other. They are 2 cisco 2950 switches, 2 Intel switches and 4 other make switches. The problem is that when I ping from my PC on Intel switch to the Cisco switch I get 400-600ms latency while when I ping any PC connected to the cisco switch I get normal >1ms latency. The CPU utilization is 9-10% on cisco switch. Pls can anybody suggest me the possible reason

 

[vipergg]

Ping response on any network switch is given low priority over other things that the switch does so the ping will always be longer than pinging an attached device . To get true response you have to ping thru the switch to an attached device . Also I would check all uplinks between your switches and make sure you don't have any speed/duplex mismatches as this will cause all kinds of errors and slow response in your network between switches.

 

[burtsbees]

From what I read, it looks like that he IS pinging through both switches...
pc---Intel---Cisco----pc=1ms
pc---Intel---Cisco=400-600ms
Is this correct?

Burt

 

[arvindai]

burtsbees this is exactly the same case, moreover when I ping from the cisco swtich to any switch or PC I get the same latency. However I have diagnosed that in the evening the problem vanishes and the cisco switches ping just fine with normal latency

 

[arvindai]

Can I create SNMP traps on my cisco switch so that I come to know what's the cause of the problem, If any PC is creating excessive traffic due to viruses or other problem.

 

[burtsbees]

You can, but I would use Ethereal or some other packet sniffer. Many in this forum have suggested Etherpeek, also. Are there any pc's with a peer-to-peer file sharing program, like Limewire or a bit torrent?

Burt

 

[BuckWeet]

Do you have any multicast running??

 

[arvindai]

How can I check to see if any multicast traffic running, pls clarify.

 

[BuckWeet]

well the best way would be to utilize a sniffer.. or do an audit on the applications running on your network..


When multicast is configured improperly it can cause network devices to process switch the traffic..


BuckWeet

 

[arvindai]

when I do a show log on the cisco switch, the output shows:

1y9w: %SPANTREE-2-RECV_PVID_ERR: Received BPDU with inconsistent peer vlan id 881 on FastEthernet0/1 VLAN1.
1y9w: %SPANTREE-2-BLOCK_PVID_LOCAL: Blocking FastEthernet0/1 on VLAN0001. Inconsistent local vlan.
1y9w: %SPANTREE-2-UNBLOCK_CONSIST_PORT: Unblocking FastEthernet0/1 on VLAN0001.Port consistency restored.

I am unable to figure out the problem, pls help

 

[lerdalt]

Is Fastethernet 0/1 your connection between the Intel and Cisco switch?

Whatever interface is the connection between the 2 make certain you have removed spanning-tree portfast from that interface.

I found this in a document from TAC:
"SPANTREE-2-BLOCK-PVID-LOCAL: Blocking [chars] on [chars] Inconsistent local vlan.

Explanation This message means that the spanning-tree port associated with the listed spanning-tree instance and interface will be held in spanning-tree blocking state until the port VLAN ID (PVID) inconsistency is resolved. The listed spanning-tree instance is that of the native VLAN ID of the listed interface. The first [chars] is the spanning-tree instance; the second [chars] is the interface.

Action Verify that the configuration of the native VLAN ID is consistent on the interfaces on each end of the 802.1Q trunk connection. Once corrected, spanning tree automatically unblocks the interfaces as appropriate. "

Check the port configuration on both sides to make certain any trunking is working correctly.

On a really basic thought...speed/duplex mismatch?

 

[dtrtel]

I agree with lerdalt, best practice is to lock down speed/duplex between switches, routers, servers--basically anything that really should not change after set up. Then disable spanning-tree portfast on all cisco switchports connected to non-cisco switches.


--jeff

http://linkedin.com/in/JeffLynch