I recently had an experience working on a couple of 6500-series switches in a VSS configuration. Both switches were running IOS 12.2(33)SXI. I found an interesting behavior when working on the switches. I’d seen this type of behavior before on lower-end switches (Cisco/IBM CIGESM Blade Center switches), but had not seen this behavior before on the 6500-series.
While working on the switch, I started to notice some errors that were being generated, namely:
*Feb 4 10:01:53.463 MST: %EC-SW1_SP-5-CANNOT_BUNDLE2: Gi1/5/2 is not compatible with Gi2/5/2 and will be suspended (trunk mode of Gi1/5/2 is dynamic, Gi2/5/2 is trunk)
SWITCH#sh run int gi1/5
*Feb 4 10:02:03.467 MST: %EC-SW1_SP-5-CANNOT_BUNDLE2: Gi1/5/2 is not compatible with Gi2/5/2 and will be suspended (trunk mode of Gi1/5/2 is dynamic, Gi2/5/2 is trunk)
Well, I had been configuring both of these switches from the ground-up, so this sparked my curiosity as I rarely (if ever) use dynamic mode for any ports. What I found was a little disconcerting, in that the running configuration really isn’t the running configuration:
Building configuration…
Current configuration : 160 bytes
!
interface GigabitEthernet1/5/2
switchport
switchport trunk encapsulation dot1q
switchport mode trunk
switchport nonegotiate
channel-group 2 mode on
end
SWITCH#sh run int gi2/5/2
Building configuration…
Current configuration : 160 bytes
!
interface GigabitEthernet2/5/2
switchport
switchport trunk encapsulation dot1q
switchport mode trunk
switchport nonegotiate
channel-group 2 mode on
end
SWITCH#
Notice that both interface configurations above show a configured mode of trunk? Yet the switch reports that one is dynamic and the other is trunk. I tried the same thing I’ve used before on the CIGESM – re-enter the commands that are already appearing in the running configuration.
*Feb 4 10:02:13.467 MST: %EC-SW1_SP-5-CANNOT_BUNDLE2: Gi1/5/2 is not compatible with Gi2/5/2 and will be suspended (trunk mode of Gi1/5/2 is dynamic, Gi2/5/2 is trunk)
SWITCH#conf t
Enter configuration commands, one per line. End with CNTL/Z.
SWITCH(config)#int gi1/5/2
*Feb 4 10:02:23.471 MST: %EC-SW1_SP-5-CANNOT_BUNDLE2: Gi1/5/2 is not compatible with Gi2/5/2 and will be suspended (trunk mode of Gi1/5/2 is dynamic, Gi2/5/2 is trunk)
SWITCH(config)#int gi1/5/2
SWITCH(config-if)#switch mode trunk
SWITCH(config-if)#end
SWITCH#
*Feb 4 10:02:37.343 MST: %SYS-5-CONFIG_I: Configured from console by cisco on console
SWITCH#
This is obviously a bug. I didn’t open a TAC case about this, nor did I look at the bug toolkit, as I found a workaround. It’ll be interesting to see if 12.2(33)SXI1 will have this fixed…
Related posts:
- NVRAM errors when saving configuration I was working on a client switch (VSS, modular IOS...
- EtherChannel behavior in NX-OS I’ve had a lot of discussions with clients about the...
- Cisco VSS presentation Here's a detailed technical presentation on Cisco VSS 1440 that...
- Beware the config register… Beware of the two configuration registers on 6500 supervisor engines/modules....
- Missing VLANs are not automatically created in NX-OS In IOS, if we assign a switchport to a non-existent...
Related posts brought to you by Yet Another Related Posts Plugin.
Tags: 12.2(33)SXI, bug, cisco, config, configuration, ios, virtual switching system, vss
I saw the same behavior in an older SX-IOS (121-27b(E4)).
There, the IOS obviously didn’t forget parts of previous definititions for interfaces.
The issue could be solved by setting the switchport purposely to “mode access” and then to “mode trunk” again.
I hope your workaround was as easy.
So, they (n)ever come back (the bugs)…
Thank you for the great feedback! If I run into this again, I’ll try changing the switchport mode and see if that resolves it. This is faster than re-pasting in the existing interface configuration (which resolved the issue when I encountered it).