forwarding on MST instance whilst in loop inconsistency on the port

[andy!doesnt!like!uucp]

Hi dudes

today i've met something really wierd on C6.5K (WS-C6506-E + WS-SUP720-3B + WS-X6708-10GE + 12.2(33)SXI3).

in substance:

#sho spanning-tree interface po4

Mst Instance Role Sts Cost Prio.Nbr Type
------------------- ---- --- --------- -------- --------------------------------
MST0 Desg FWD 1000 128.1667 P2p Bound(PVST)
MST1 Desg FWD 1000 128.1667 P2p Bound(PVST) *LOOP_Inc

Isnt it funny? i made diagnosis from number of perspectives & identified no deviations except above one.

no bug-search helped as well. any idea?

[Mohsin Alam]

looks like a UDLD failure, did you check L1?

Have you configured loop guard as well?

If yes can you post the output of show span de | sec port-channel 4

## Make sure to mark post as helpful, If it resolved your issue. ##

[andy!doesnt!like!uucp]

Hi

as i said no error on L1 is visible:

Group Port-channel Protocol Ports
------+-------------+-----------+-----------------------------------------------
4 Po4(SU) LACP Te6/2(P) Te6/4(P)

Port Device Name Device ID Port ID Neighbor State
---- ----------- --------- ------- --------------
Te6/2 SSI134701Y0 1 Ethernet1/1 Bidirectional
Te6/4 SSI134701Y0 1 Ethernet1/5 Bidirectional

loop guard is configured globally. & the mst 1 meets loop_inc prerequisite to trigger by no BPDU ever received:

Port 1667 (Port-channel4) of MST1 is forwarding (Loop Inconsistent)
Port path cost 1000, Port priority 128, Port Identifier 128.1667.
Designated root has priority 8193, address 001d.e604.9d40
Designated bridge has priority 16385, address 001d.e58b.f0c0
Designated port id is 128.1667, designated path cost 2500
Timers: message age 0, forward delay 0, hold 0
Number of transitions to forwarding state: 4
Link type is point-to-point by default, Boundary PVST
PVST Simulation is enabled by default
Loop guard is enabled by default on the port
BPDU: sent 8077699, received 0

 & that's all but port is still in forwarding state for mst1.

i got lost. 

[Mohsin Alam]

Can you check the device connected to PO4, any high cpu as such or abnormalities? why is it no sending BPDU's?

Also post the output of show int po4 

## Make sure to mark post as helpful, If it resolved your issue. ##

[andy!doesnt!like!uucp]

at the moment i cant check opposite switch, but 0 BPDU counters on RX of Po4 is not enough to bring instance in loop inconsistent state.

moreover in Loop inconsistent instance cannot be in FWD state, or i'm wrong?

not sure how below output can be helpful:

Port-channel4 is up, line protocol is up (connected)
Hardware is EtherChannel, address is 001a.6c9e.8001 (bia 001a.6c9e.8001)
Description: --> nexus
MTU 1500 bytes, BW 20000000 Kbit, DLY 10 usec,
reliability 255/255, txload 1/255, rxload 1/255
Encapsulation ARPA, loopback not set
Keepalive set (10 sec)
Full-duplex, 10Gb/s, media type is unknown
input flow-control is on, output flow-control is off
Members in this channel: Te6/2 Te6/4
ARP type: ARPA, ARP Timeout 04:00:00
Last input never, output never, output hang never
Last clearing of "show interface" counters 1y21w
Input queue: 0/2000/0/0 (size/max/drops/flushes); Total output drops: 0
Queueing strategy: fifo
Output queue: 0/40 (size/max)
5 minute input rate 25996000 bits/sec, 3527 packets/sec
5 minute output rate 7826000 bits/sec, 2374 packets/sec
223907544752 packets input, 176470558633443 bytes, 0 no buffer
Received 202518743 broadcasts (110613522 multicasts)
0 runts, 0 giants, 0 throttles
7066 input errors, 0 CRC, 0 frame, 0 overrun, 0 ignored
0 watchdog, 0 multicast, 0 pause input
0 input packets with dribble condition detected
222424938626 packets output, 132456340507886 bytes, 0 underruns
0 output errors, 0 collisions, 0 interface resets
0 babbles, 0 late collision, 0 deferred
0 lost carrier, 0 no carrier, 0 PAUSE output
0 output buffer failures, 0 output buffers swapped out

[Mohsin Alam]

You are right, looks like a buggy behavior 

Also i see input errors :

7066 input errors, 0 CRC, 0 frame, 0 overrun, 0 ignored

Are they historical? 

## Make sure to mark post as helpful, If it resolved your issue. ##

[andy!doesnt!like!uucp]

Hi

i've got Insufficient Permissions to View Bug with my account bound to my company (

errors r historical

[Giuseppe Larosa]

Hello @andy!doesnt!like!uucp ,

this bug is visible only to Cisco employees I am too not able to see it

Hope to help

Giuseppe

[Mohsin Alam]

If the input errors are historical then it looks good, just need to check on neighboring device.
the bug basically says that loop inc can be in FWD state due to misconfig of loop guard and MST 

## Make sure to mark post as helpful, If it resolved your issue. ##

[andy!doesnt!like!uucp]

could u pls keep posting with more details on bug?

[andy!doesnt!like!uucp]

@Mohsin Alam so could u pls advice on how to clear this state?

[Giuseppe Larosa]

Hello @andy!doesnt!like!uucp ,

a possible workaround is to change configuration of root bridge for MST 1 so that you solve the loop inconsistency.

i.e make root bridge for MST 1 the same device that is root bridge for MST 0 this should solve your issue. I know that this way you miss the diverse topology per MST.

Or in the opposite knowing that a single MST PDU is actually sent for all MSTs you can cause the loop inconsistency also on MST 0 at that point likely the port will become loop inco for all MST instances and it may become not forwarding for all MST instances.

CAUTION: here this is just a guess that the port follows what MST 0 decides ( for the SW bug)  so the second way is unsafe.

Hope to help

Giuseppe

[andy!doesnt!like!uucp]

Hi Giuseppe

i'm not sure that MST 1 is in really blocking state on Po4 (notice switch reports it's forwarding). I also wouldnt like to change root for MST 1 before i fully learn this deployment (it came to me quite recently). Currently root for MST 1 is on other C6.5K & i dont want to change it unless i have solid base to do this. i'd like to clear LOOP_INC state w/o reconvergency on MST 1. But looks like there is no option for this. i'm thinking about artificial flapping Po4 from either side to clear LOOP_INC. 

[Mohsin Alam]

You will need to check on the neighboring device which is NEXUS, for removing the Loop incs* you can disable loop guard ( which i don't think you will like to do ) 

the main issue here is why is NEXUS not sending BPDU's .
The second part of the PO4 being in forwarding state even after loop incs* 

Also can you share the STP config of both the devices if you have it handy. 

By any chance is BPDU filter applied? 

## Make sure to mark post as helpful, If it resolved your issue. ##

[andy!doesnt!like!uucp]

Hi @Mohsin Alam 

if i understand your 1st statement correctly temporary disablement of loopguard shell remove LOOP_INCs. Good idea.

C6.5K doesnt show any BPDUs from N5K because of N5K runs RPVST+. Isnt it expected behavior?

STP config is pretty simple on each side & no filtering nor other deviations r visible.

C6.5K

spanning-tree mode mst
spanning-tree loopguard default
spanning-tree pathcost method long
spanning-tree mst configuration

name MSTP
revision 1
instance 1 vlan 14-19, 256

spanning-tree mst 0 priority 8192
spanning-tree mst 1 priority 16384

N5K

spanning-tree mode rapid-pvst
spanning-tree pathcost method long
spanning-tree bridge assurance
spanning-tree vlan 1,4-5,11-23,65,70-73,80,96,110,112,120,128,192,256-258,301-308,310-311,324,363,400,800-807,820-826,902,990,999
spanning-tree vlan 1,4-5,11-23,65,70-73,80,96,110,112,120,128,192,256-258,301-308,310-311,324,363,400,800-807,820-826,902,990,999 priority 32768
spanning-tree vlan 1,4-5,11-23,65,70-73,80,96,110,112,120,128,192,256-258,301-308,310-311,324,363,400,800-807,820-826,902,990,999 hello-time 2
spanning-tree vlan 1,4-5,11-23,65,70-73,80,96,110,112,120,128,192,256-258,301-308,310-311,324,363,400,800-807,820-826,902,990,999 forward-time 15
spanning-tree vlan 1,4-5,11-23,65,70-73,80,96,110,112,120,128,192,256-258,301-308,310-311,324,363,400,800-807,820-826,902,990,999 max-age 20