Hey everyone,

I know, GNS3 is not best suited for handling switches, but some features are available, and HSRP is one of them.

In my topology https://docs.google.com/open?id=0B6XxNd5c3zV_bGV2a09lZ0JwaWs made up of 3725s with NM-16ESW cards, if one of the VPCS "hosts" tries to ping one of the Virtual IP addresses, it fails. The pings attain the right active router, but this one never answers.

Any suggestion?

For those of you who want to test this lab out within GNS3: https://docs.google.com/folder/d/0B6XxNd5c3zV_cGFnX3dxWnpGRFE/edit

Edited on 04/20/2015: Solved in GNS3/Linux 1.3.1

_________________
A Collection of GNS3 Labs: https://learningnetwork.cisco.com/message/232820
i7 4700MQ quad-core 2.6 Ghz and 16 GB RAM
CCNA - CCDA
Beta Test Team
GNS3 0.8.6 on Windows 7 64 bits
GNS3 0.8.6 on Ubuntu 13.10 and Qemu 1.5.2
@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@

Hey, I've found the *glitch* and it's weird.

For an unknown reason, the following commands are *automatically added by GNS3* each time you define an HSRP group on a SVI with a non default priority:*
"mac-address-table static 0000.0c07.acxx interface FastEthernet1/0 vlan n" where xx is replaced by the group number and n by the vlan number.

Basically, they mean: you can access the "HSRP group xx" MAC addresses on interface Fa0/1, whereas these MAC addresses are managed by the local switch (in my config, the only priorities defined are the biggest of the group)!!!

All these commands are useless; you need to clear them with "#clear mac-address-table static" and *HSRP works fine on both switches*! :)

It would be even better if this glitch was corrected someday ;)

_________________
A Collection of GNS3 Labs: https://learningnetwork.cisco.com/message/232820
i7 4700MQ quad-core 2.6 Ghz and 16 GB RAM
CCNA - CCDA
Beta Test Team
GNS3 0.8.6 on Windows 7 64 bits
GNS3 0.8.6 on Ubuntu 13.10 and Qemu 1.5.2
@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@

Correction: "Automatically added by IOS". GNS3 has nothing to do in that.

Cheers,

_________________
Jeremy, GNS3 Programmer & Benevolent Dictator for Life.

Good catch actionmystique! I had spent quite some time on trying to figure this one out.

But this is really weird!

Firstly, you wouldn't think the static mac address would make any difference - but clearly it does. I'm convinced this is a bug in the IOS - but would love someone who has real hardware to be able to verify it.

Secondly, if you remove the <b>mac-address-table static</b> lines from your config, save your config and reload your router - the stupid things come back!

Anyway, I'm glad someone solved it.

_________________
RedNectar
http://rednectar.net
@rednectarchris
GNS3 WorkBench-a VMware image of Ubuntu with GNS3 and VPCS installed and a collection of exercises/labs

Hi
I also experienced this problem - And I still don't understand why this happen. But maby someone can use the following to come up with an idea:

I made a default HSRP setup with four 3660 routers and NM-16ESW module. Two DLS and two ALS
IOS: Version 12.4(15)T1
DLS1 connected to DLS2 via fa 1/10
DLS1 connected to ALS1 via fa 1/6 and to ALS2 via 1/8
DLS2 connected to ALS2 via fa 1/6 and to ALS1 via 1/8

Summary:
When I test the HSRP function by shutting down the SVI on the active router and then no shutdown, everything functions fine. I've never seen the problem in this situation.
But sometimes when booting the system or rebooting only DLS1 something goes wrong. I get a static entry for the virtual mac address in the mac-address-table pointing to the standby router/switch.
DSL1#sh mac-address-table
Destination Address Address Type VLAN Destination Port
------------------- ------------ ---- --------------------
0000.0c07.ac01 Static 1 FastEthernet1/10
cc01.1bb0.0000 Dynamic 1 FastEthernet1/10

I only happens sometimes (and I tried many times) - therefore I suspect it to be a timing problem - maby because it goes to fast in the virtual system compared to the physical system.
I have tried with debug standby on both DLSs but the output seems equal in all tests.
I also tried to add the delay parameter to the preempt function - but it seems that the delay function does not function when rebooting.

I don't have a NM-16ESW module physically so I'm not able to compare with the virtual environment.

Once I did see the virtual mac-address type to be Self and then it changed to Static - doing sh mac-address-tabel all the time during the switchover.

I hope that someone is able to explain why this happen

Documentation:
on SVI interface:
(DLS1)
...
int vlan 1
ip address 172.16.1.2 255.255.255.0
standby 1 ip 172.16.1.1
standby 1 priority 150
standby 1 preempt

(DLS2)
...
int vlan 1
ip address 172.16.1.3 255.255.255.0
standby 1 ip 172.16.1.1
[standby 1 priority 100] ! not nessesary because it is default
standby 1 preempt

----------------------------------------------------------------------------
----------------------------------------------------------------------------
! Test:
! DLS1 is Active and DLS2 is standby
----------------------------------------------------------------------------
----------------------------------------------------------------------------
DSL1#sh standby brief
P indicates configured to preempt.
|
Interface Grp Pri P State Active Standby Virtual IP
Vl1 1 150 P Active local 172.16.1.3 172.16.1.1

DSL1#sh mac-address-table
Destination Address Address Type VLAN Destination Port
------------------- ------------ ---- --------------------
cc02.1ff0.0000 Self 1 Vlan1
cc01.1ff0.0000 Dynamic 1 FastEthernet1/10

DSL2#sh mac-address-table
Destination Address Address Type VLAN Destination Port
------------------- ------------ ---- --------------------
cc01.1ff0.0000 Self 1 Vlan1
0000.0c07.ac01 Dynamic 1 FastEthernet1/10

----------------------------------------------------------------------------
! Shutdown of SVI on DLS1:
----------------------------------------------------------------------------
int vlan 1
shutdown

then DLS2 becomes Active - that's fine
DSL1#sh mac-address-table
Destination Address Address Type VLAN Destination Port
------------------- ------------ ---- --------------------
0000.0c07.ac01 Dynamic 1 FastEthernet1/10
cc02.1ff0.0000 Self 1 Vlan1

DSL2#sh mac-address-table
Destination Address Address Type VLAN Destination Port
------------------- ------------ ---- --------------------
cc01.1ff0.0000 Self 1 Vlan1

----------------------------------------------------------------------------
! no shutdown of SVI on DLS1:
! DLS1 becomes Active - that's fine
----------------------------------------------------------------------------

----------------------------------------------------------------------------
! shutdown of fa 1/6, 1/8 and 1/10 (all interfaces to neighbour switches)
----------------------------------------------------------------------------
DLS2 becomes Active - that's fine
Then no shutdown and DLS1 again becomes active - fine but
sometimes you get a static entry for the virtual mac address!

DSL1#sh mac-address-table
Destination Address Address Type VLAN Destination Port
------------------- ------------ ---- --------------------
0000.0c07.ac01 Static 1 FastEthernet1/10
cc01.1bb0.0000 Dynamic 1 FastEthernet1/10

DSL2#sh mac-address-table
Destination Address Address Type VLAN Destination Port
------------------- ------------ ---- --------------------
cc01.1bb0.0000 Self 1 Vlan1
0000.0c07.ac01 Dynamic 1 FastEthernet1/10

How long have you waited for it to happen ?

_________________
A Collection of GNS3 Labs: https://learningnetwork.cisco.com/message/232820
i7 4700MQ quad-core 2.6 Ghz and 16 GB RAM
CCNA - CCDA
Beta Test Team
GNS3 0.8.6 on Windows 7 64 bits
GNS3 0.8.6 on Ubuntu 13.10 and Qemu 1.5.2
@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@

I've just tested it. One primary and one standby 3725s with NM-16ESW module.
Started 2 other 3725s without HSRP implemented and instantly boom: one wrong static MAC address in the MAC table on primaries (SW1 for group 1 and SW2 for group 4):


... with this common topology:

_________________
A Collection of GNS3 Labs: https://learningnetwork.cisco.com/message/232820
i7 4700MQ quad-core 2.6 Ghz and 16 GB RAM
CCNA - CCDA
Beta Test Team
GNS3 0.8.6 on Windows 7 64 bits
GNS3 0.8.6 on Ubuntu 13.10 and Qemu 1.5.2
@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@

I forgot to mention in the previous topology that SW1 and SW2 needs another switch to be able to send each other HSRP multicast, because their direct connection is based on L3 routed ports.

This issue might be linked to preemption: before both switches can send each other HSRP multicast, nothing happens, after they are connected at L2 with SW4 switched on, the wrong MAC entries in the MAC address tables appear:

_________________
A Collection of GNS3 Labs: https://learningnetwork.cisco.com/message/232820
i7 4700MQ quad-core 2.6 Ghz and 16 GB RAM
CCNA - CCDA
Beta Test Team
GNS3 0.8.6 on Windows 7 64 bits
GNS3 0.8.6 on Ubuntu 13.10 and Qemu 1.5.2
@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@

Hi
Hmm - we talked about (my students and I) and we did only see the problem with the static mac-address if there was a change in the Active/Standby roles. And we did only see it on DLS1.

The change in Active/Standby could be when booting the topology
or
If the host pc was so busy that there was a loss in hellos between the two HSRP routers.

Maby it was the last thing that happend when you started two additional routers - I don't know

Just for you to know, maybe it helps.

I configured HSRP on two 3660 with c3660-is-mz.124-10a.bin IOS
No problems at all (up to now...)

Maybe that one has no bug