MAC Address Problem on SGE2010

[Lodac1977]

I have 10 SGE2010 switches. There is a stack of 4 PoE switches that is the L3 core. There is another stack of 4 data only switches. And another stack of 2 PoE switches. The 4 and 2 switch stacks connect back to the 4 switch core stack via single ethernet connection.

I am having an issue where users will undock their laptops to go to a conference room, plug in at the conference room and be fine. Then, they go back to their docking station and plug in and have no network connectivity. I have to go in and clear all of the MAC Address tables on all the stacks in order for their network connectivity to be restored. This is all via wired connections.

There is no port security of any kind enabled. No ARP inspection, snooping, etc.

Aging timers on ARP and MAC are the defaults for the switches. 60,000 and 300 respectively.

Any ideas what could be causing this and how to fix it so that moving around the campus doesn't cause loss of connectivity?

Firmware version is 3.0.0.18  Boot Version is 2.0.0.03  Hardware Version is 00.00.01

[kalp pan]

Hi I have a similar problem with these same switches and the firmware versions you mention.

Basically a wired person can hit intervlans and their gateway.

so host ip 192.168.11.33   can hit gw 192.168.11.1

It can hit 192.168.252.1 which is the gw ip for the routing engine on the switch.

But it cant hit 192.168.252.6 which is the default gateway out.

This doesnt affect all users, it only affects some users when they switch networks around.

I find it really strange that one machine behaves this way, then about 2-3 mins later it starts working fine again, I can ping the default gateway out 192.168.252.6

All this while other users are able to ping 192.168.252.6 just fine when someone else has this problem.

It is random and happens to random people in the office.

I am about to upgrade the firmware to the lastest revision, but I dont think it will fix anything.

[rmanthey]

Kal,

We might need to get more information, but I want to confirm a few things. This is also for Lodac, if it does start working again, try to see if it is the default age out timer of 5 min. The MAC address will not be learned any where else until it is aged out in the port it was originally in if you are using port security. Lodac you stated you are not using port security, is that on all switches or just one of the three stacks? If you leave yours alone for 5 min. will it start communicating again lodac?

We might need for the both of you to call in to the 1-866-606-1866 center so we can get more detailed information about your networks.

Cisco Small Business Support Center

Randy Manthey

CCNA, CCNA - Security

[kalp pan]

For me the switch is in L3 mode.  It is routing 3 diff lans.

192.168.8.0/22  users

192.168.12.0/22 users

192.168.252.0/24 gateways

The default gateway out is 192.168.252.6, the routing engine ip addresses are .1 on those vlans.

When this problem occurs with some user, he can reach everthing in .8 and .12 including 192.168.252.1

He just cant reach 192.168.252.6.

All other users who arent facing the issue can ping 192.168.252.6 just fine.  I have tried switching the default gateway to 2 different devices, they both behave the same.  (I have a SRX 220 192.168.252.2 and ASA 5520 192.168.252.6 as gateways with different isps for failover manually).

I thought it was the aging problem so I have even changed the mac dynamic aging timeout to 10s,  So the user should get going after 10s, but it doesnt work.   The duration remains the same.

I have even changed the arp aging timeout to 20s.  Still randomly this continues to happen.

It has even happened to myself when I have not switched any network recently, but I can no longer ping 252.6

I have verified there all negotiations are good and no interface errors or resets are happening.

I will ugprade the firmware this weekend for the switches, I will call in after that if this problem persists.

Thanks for helping.

[Peter Marquis]

I am facing a similar issue with one of my clients, Every so often, a host loses connectivity to some services/part of the network and the only way to cure it is to clear ther arp table on the SGE stack they are connected to.

     Been hunting the web for an answer, but nothing is forthcomming. It was happening a while back and a firmware upgrade seemed to fix it, but it has come back. Anyone any ideas?