Cisco peeps...

[wildblue 4]

I have a 3550 that seems to have forgotten how to route some addresses. It first started with just one subnet - a machine wouldn't be able to get off its own subnet (ok, traffic was actually getting off the subnet, the switch just wouldn't route the return traffic) but just changing the machine's IP address would make it start to work again. Seems to be about one machine a day it'll happen to. Happened with the voice vlan too (you could hear them talking on the phone, they couldn't hear you)

Narrowed the problem down to one of the core switches that's doing the routing between VLANs (each subnet on its own vlan) -- there's two core switches, x.x.x.2 and x.x.x.3 with a standby address of x.x.x.1. If I change my gateway to point right at the 'failover' router x.x.x.3, everything works fine. Short of booting the x.x.x.2 switch, any ideas?

it's like incest - you're substituting convenience for quality

[normiss 622]

misconfigured HSRP?

buggy code? I almost always start with updating the code...and if you have Cisco Works in place, it's painless.

[wildblue 4]

It is an older IOS, probably needs updating. There's just never a good time to do it to this one. No Cisco Works either ... been trying to get it approved though. Was just hoping for a quick easy fix to this one.

it's like incest - you're substituting convenience for quality

[ryoder 1,393]

Quote

It is an older IOS, probably needs updating.

Contrary to popular myth, code doesn't develop new bugs as it ages.
Do "show ip route" to see what routes each unit knows, and "show standby" to see the state of HSRP on each unit.

"There are only three things of value: younger women, faster airplanes, and bigger crocodiles" - Arthur Jones.

[normiss 622]

took you long enough to get here...

care to do some VoIP troubleshooting at the same time?

I'm having dial-peer and translation issues as the traffic passes through the Juniper ATM cloud over to the Marconi M320's....not including their refusal to pass pref bits...

[ryoder 1,393]

Quote

took you long enough to get here...

care to do some VoIP troubleshooting at the same time?

I'm having dial-peer and translation issues as the traffic passes through the Juniper ATM cloud over to the Marconi M320's....not including their refusal to pass pref bits...

I haven't worked with VOIP. I've considered getting some VOIP modules for my routers to play with it, but haven't had the time.

"There are only three things of value: younger women, faster airplanes, and bigger crocodiles" - Arthur Jones.

[wildblue 4]

No, but maybe this is a bug that was fixed with a later release that we're just now seeing for some reason. I dunno ... does seem odd.

routes are both the same - it's just routing directly connected VLANs, one goofy static route and a gateway of last resort.

The x.x.x.2 is the active router, .3 is the standby, x.x.x.1 - they both agree on this

Local state is Active, priority 100, may preempt
Local state is Standby, priority 100, may preempt

Should I just change priority on the standby?

it's like incest - you're substituting convenience for quality

[ryoder 1,393]

Quote

Local state is Active, priority 100, may preempt
Local state is Standby, priority 100, may preempt

It is odd to have preemption enabled when the priority is the same. It is meaningless in that situation. My suspicion would be that the active unit is periodically losing a route or an interface is flapping. Is there a dynamic routing protocol in use?

"There are only three things of value: younger women, faster airplanes, and bigger crocodiles" - Arthur Jones.

[DaMan 0]

Quote

No, but maybe this is a bug that was fixed with a later release that we're just now seeing for some reason. I dunno ... does seem odd.

routes are both the same - it's just routing directly connected VLANs, one goofy static route and a gateway of last resort.

The x.x.x.2 is the active router, .3 is the standby, x.x.x.1 - they both agree on this

Local state is Active, priority 100, may preempt
Local state is Standby, priority 100, may preempt

Should I just change priority on the standby?

Yes - change it!

Also, make sure you have setup a different standby group number for each VLAN:
!
!SW.2
interface Vlan5
standby 5 ip x.x.x.1
standby 5 priority 120
standby 5 preempt
!
!SW.3
interface Vlan5
standby 5 ip x.x.x.1
standby 5 priority 100
standby 5 preempt

!
!SW.2
interface Vlan6
standby 6 ip x.x.x.1
standby 6 priority 120
standby 6 preempt
!
!SW.3
interface Vlan6
standby 6 ip x.x.x.1
standby 6 priority 100
standby 6 preempt

Good luck!

Z-Flock 8
Discotec Rodriguez

Too bad weapons grade stupidity doesn't lead to sterility.

[wildblue 4]

OSPF is configured, but it's not actually distributing anything (its only neighbor is the other core switch, all routes are "C" or "S" ... no "O") ... not even sure why it's configured. Maybe was in use at one time.

I'm guessing the priority is the same so whoever set it up could just copy and paste and let the lower IP have priority.

I do have different standby groups for each interface.

I might just give the other router a higher priority to fix the problem now. I don't think the active one is losing a route or has a flapping interface - it routes for 30 other devices on the same subnet just fine. Ex: One machine has a DHCP assigned address of 192.168.10.123 - it stops working, but as soon as I manually change the address to 192.168.10.121 it works fine. The machine next to it can have an address of x.x.10.124 and it'll work fine. It's starting to periodically affect the phones in that one area as well (different VLAN/subnet)

EDIT: The standby priority isn't actually set, the config reads
standby xx ip x.x.x.x
standby xx preempt

the 'priority 100' only shows up in the 'show standby' command, so I'll assume that's just the default.

it's like incest - you're substituting convenience for quality