Re: Loss of management

k.a.clarke · ‎07-20-2006

Hi

I keep losing connectivity to the management IP address on 2950 switches, although all other VLANs continue to work through it.

When I plug in a console connection, loads of buffered messages are output (just port up/down messages), and then IP starts working again.

This is happening on lots (but not all) of switches in lots of different locations.

Is anyone else aware of such a bug?

All the switches are running: 121-22(EA6)

A sample of the console messages...

Jul 5 19:02:22: %LINK-CLUSTER_MEMBER_1-3-UPDOWN: Interface FastEthernet0/3, cha

nged state to down

Jul 5 19:02:23: %LINK-CLUSTER_MEMBER_1-3-UPDOWN: Interface FastEthernet0/3, cha

nged state to up

Jul 5 19:07:29: %LINK-CLUSTER_MEMBER_1-3-UPDOWN: Interface FastEthernet0/4, cha

nged state to down

Jul 5 19:07:30: %LINK-CLUSTER_MEMBER_1-3-UPDOWN: Interface FastEthernet0/4, cha

nged state to up

Jul 5 19:07:55: %LINK-CLUSTER_MEMBER_1-3-UPDOWN: Interface FastEthernet0/3, cha

nged state to down

Jul 5 19:07:

Jul 11 00:37:07: %LINK-CLUSTER_MEMBER_1-3-UPDOWN: Interface FastEthernet0/2, cha

nged state to down

Jul 11 00:37:09: %LINK-CLUSTER_MEMBER_1-3-UPDOWN: Interface FastEthernet0/2, cha

nged state to up

Jul 11 00:40:54: %LINK-CLUSTER_MEMBER_1-3-UPDOWN: Interface FastEthernet0/13, ch

anged state to down

Jul 11 00:40:56: %LINK-CLUSTER_MEMBER_1-3-UPDOWN: Interface FastEthernet0/13, ch

anged state to up

Jul 11 00:40:58: %LINK-CLUSTER_MEMBER_1-3-UPDOWN: Interface FastEthernet0/9, cha

nged state to down

Jul 11 00:41:00: %LINK-CLUSTER_MEMBER_1-3-UPDOWN: Interface FastEthernet0/9, cha

nged state to up

Jul 11 00:44:06: %LINK-CLUSTER_MEMBER_1-3-UPDOWN: Interface FastEthernet0/13, ch

anged state to down

Jul 11 00:44:07: %LINK-CLUSTER_MEMBER_1-3-UPDOWN: Interface FastEthernet0/13, ch

anged state to up

Jul 11 00:46:21: %LINK-CLUSTER_MEMBER_1-3-UPDOWN: Interface FastEthernet0/13, ch

anged state to down

Jul 11 00:46:23: %LINK-CLUSTER_MEMBER_1-3-UPDOWN: Interface FastEthernet0/13, ch

anged state to up

Jul 11 00:46:37: %LINK-CLUSTER_MEMBER_1-3-UPDOWN: Interface FastEthernet0/2, cha

nged state to down

Jul 11 00:46:39: %LINK-CLUSTER_MEMBER_1-3-UPDOWN: Interface FastEthernet0/2, cha

nged state to up

Jul 11 00:48:44: %LINK-CLUSTER_MEMBER_1-3-UPDOWN: Interface FastEthernet0/13, ch

anged state to down

Jul 11 00:48:45: %LINK-CLUSTER_MEMBER_1-3-UPDOWN: Interface FastEthernet0/13, ch

anged state to up

Jul 11 00:51:06: %LINK-CLUSTER_MEMBER_1-3-UPDOWN: Interface FastEthernet0/13, ch

anged state to down

Jul 11 00:51:08: %LINK-CLUSTER_MEMBER_1-3-UPDOWN: Interface FastEthernet0/13, ch

anged state to up

Jul 11 00:54:15: %LINK-CLUSTER_MEMBER_1-3-UPDOWN: Interface FastEthernet0/2, cha

nged state to down

Jul 11 00:54:17: %LINK-CLUSTER_MEMBER_1-3-UPDOWN: Interface FastEthernet0/2, cha

nged state to up

Jul 11 01:01:08: %LINK-CLUSTER_MEMBER_1-3-UPDOWN: Interface FastEthernet0/13, ch

anged state to down

Jul 11 01:01:10: %LINK-CLUSTER_MEMBER_1-3-UPDOWN: Interface FastEthernet0/13, ch

anged state to up

Jul 11 01:07:33: %LINK-CLUSTER_MEMBER_1-3-UPDOWN: Interface FastEthernet0/13, ch

anged state to down

Jul 11 01:07:34: %LINK-CLUSTER_MEMBER_1-3-UPDOWN: Interface FastEthernet0/13, ch

anged state to up

Jul 11 01:17:11: %LINK-CLUSTER_MEMBER_1-3-UPDOWN: Interface FastEthernet0/13, ch

anged state to down

Jul 11 01:17:12: %LINK-CLUSTER_MEMBER_1-3-UPDOWN: Interface FastEthernet0/13, ch

anged state to up

Please find attached sample config.

Any help appreciated.

mbuillon · ‎07-20-2006

Hi,

you have enable prunning int the vtp domain ?

If prunning is enable and no physical port in your management vlan are actif, the paquets do not passe on the trunck... i think.

k.a.clarke · ‎07-20-2006

Hi

Thanks for your response.

No pruning is not enabled on the domain, also, switches below affected switches are able to communicate on the management vlan, eg: where switch 1s management interface is not contactable but switch 2s is.

root (6509)

|

switch 1

|

switch 2

glen.grant · ‎07-20-2006

Have not seen this on our 2950's in a trunked net , and we are using the same code. You appear to be trying to cluster these switches so that is something we don't do . Have you checked these 2 interfaces for errors and or physical type problems , they are obviously bouncing . Verify interface settings are the same on both sides ie speed and duplex and trunk settings .

k.a.clarke · ‎07-21-2006

Hi

Thanks for the response.

It's true this switch is part of a cluster (historical) but this problem is occurring on both clustered and non clustered switches. I've checked both sides of the trunk for config/physical problems and they're both fine. All access ports are fine too.

When I loose a switch I'm able to clear the arp from it's router, then the mac is re learnt and is tracable right down to the link port poss 3/4 hops, but I'm just not able to talk IP to the management address. CDP works fine. As previously stated all user vlans are not affected.

ankbhasi · ‎09-22-2006

Hi K,

I am aware of the bug on 2950 where it looses the management connectivity.

The bug is been detailed as once the switch has been up and running for a while

it looses ip management where the user can no longer ping, telnet, etc.

Users continue to pass traffic during this time and doesn't show to be

affected.

And the version you are running is affe cted with this bug

Check this bug

CSCdx26212&

HTH

Ankur

* Pls rate all helpfull post.

k.a.clarke · ‎10-12-2006

Thanks for the info Ankur.

donwood1 · ‎09-22-2006

I too am having the same problem with Vlan 1. None of the switches are clustered. I lose the ability to ping or telnet to a switch. The only thing that "fixes" is is a power reboot. All user vlans continue to work. This happens on 2900, 3500 switches in my network.