Cisco Issue

Benny · March 09, 2007, 09:45:24 AM

%MAC_MOVE-SP-4-NOTIF

we seem to have a flapping issue, MAC's being learnt then relearnt on other ports. There are no loops, intermittent loss, no pattern to failure.

Any ideas, other than the obvious appreciated. (I'm not a user so please don't suggest any really basic stuff......I'm kinda hoping Tuggs is up and about)

8 23:36:22 GMT: %MAC_MOVE-SP-4-NOTIF: Host 0000.5e00.0143 in vlan xxx is flapping between port Gi4/9 and port Gi4.

vlan xxx isn't configured on that port.....

TeaLeaf · March 09, 2007, 08:48:32 PM

I send Tugs a prompt............ no guarantees, but hopefully he'll be around to take a peek.

TL.

Benny · March 10, 2007, 01:36:07 PM

Thanks TL.

It's an odd one. We've shut down one side and clients are stable, but the symptoms are still there.

Benny · March 15, 2007, 10:31:18 PM

Sorted, nothing to see here. Move along. Bumped for the whoreness of it.

Penfold · March 15, 2007, 11:39:35 PM

You're so pushing for that 3,000th post aren't you :narnar:

PEN

Carr0t · March 16, 2007, 12:12:27 PM

Quote from: Benny;181333Sorted, nothing to see here. Move along. Bumped for the whoreness of it.

I'm curious though. We run a fair whack of Ciscos here, edge devices rather than core stuff though, but all the ideas I could come up with I figured would come under the heading of fairly basic stuff that someone might have done as a configuration cockup or whatever. So what was the problem?

Benny · March 21, 2007, 12:36:01 PM

Bug issue on code level between the core cisco and the extreme edge device which seems to be triggered by Alteon VIP traffic....

a blinder then....

Carr0t · March 21, 2007, 02:50:37 PM

Hehe, we appear to be running pretty much the inverse. We have Ciscos at the edge (though we are in the process of phasing them out for HPs) with Extremes in the core (where we aggregate connections using switches, at any rate. We use Foundries and Junipers where we need full-on routers).

Benny · March 21, 2007, 09:24:58 PM

oooo, get away from the Extremes, we had Alpines running core for 3 years and they caused us immense pain. Admittedly so did the Summits on the edge.

Now back on a Cisco core in one center, out to Cisco edges, M20's and Foundries on the other data centers and drain points.

Carr0t · March 22, 2007, 12:35:06 PM

yeah, our core at the moment is 2 BlackDiamond 10k's, and they do cause a lot of issues. They were got a while back though. We're in the process of looking for their replacements, and may well be going with something from Foundry.