Problem with STP on Comware-based A5500, 1910, 1920 -series switches

[SergeyVl87]

I have a new network based on two A5500-24G-4SFP-HI as core switches, two 1920-48G access switches (for computers and printers) and six 1910-24-PoE+ access switches (for phones and ip cameras).

Network topology and configuration fragments attached.

Switches are connected using UTP cat6 cables; there are no errors or packet lost on interfaces.

Periodically (every few days) problems with STP occurs.

STP role of a port changes for a few seconds (Root->Desi->Root).

[Rack05-PoeSw4]dis stp history
 --------------- STP slot 1 history trace ---------------
 -------------------  Instance 0   ---------------------
 Port GigabitEthernet1/0/26
   Role change   : Desi->Root
   Time          : 2017/01/11 05:31:51
   Port priority : 0.7848-596b-92e6 2001 32768.d894-0346-5c12 0
                   32768.d894-0346-5762 128.25
 Port GigabitEthernet1/0/26
   Role change   : Root->Desi (Aged)
   Time          : 2017/01/11 05:31:09
   Port priority : 0.7848-596b-92e6 2001 32768.d894-0346-5c12 0
                   32768.d894-0346-5762 128.25

 Syslog message from this switch:

Jan 11 05:31:09 Rack05-PoeSw4 %%10MSTP/5/MSTP_BPDU_RECEIVE_EXPIRY(l): Instance 0's port GigabitEthernet1/0/26
received no BPDU within the rcvdInfoWhile interval. Information of the port aged out.

 On a neighbor switch:

[Rack05-PoeSw3]dis stp history
 --------------- STP slot 1 history trace ---------------
 -------------------  Instance 0   ---------------------
 Port GigabitEthernet1/0/26
   Role change   : Desi->Root
   Time          : 2017/01/11 05:31:51
   Port priority : 0.7848-596b-92e6 3001 32768.d894-0346-5aaa 0
                   32768.d894-0346-5c12 128.25
 Port GigabitEthernet1/0/25
   Role change   : Root->Alte
   Time          : 2017/01/11 05:31:51
   Port priority : 0.7848-596b-92e6 4000 32768.d894-0346-5aaa 0
                   32768.d894-0346-5ac8 128.26
 Port GigabitEthernet1/0/26
   Role change   : Root->Desi
   Time          : 2017/01/11 05:31:10
   Port priority : 32768.d894-0346-5c12 0 32768.d894-0346-5aaa 0
                   32768.d894-0346-5c12 128.25
 Port GigabitEthernet1/0/25
   Role change   : Alte->Root
   Time          : 2017/01/11 05:31:10
   Port priority : 0.7848-596b-92e6 4000 32768.d894-0346-5aaa 0
                   32768.d894-0346-5ac8 128.26

 Syslog from CoreSw1:

Jan 11 05:31:10 CoreSw1 %%10MSTP/6/MSTP_NOTIFIED_TC(l): Instance 0's port GigabitEthernet1/0/21 was notified
of a topology change.
Jan 11 05:31:10 CoreSw1 %%10ARP/5/ARP_DUPLICATE_IPADDR_DETECT(l): Detected an IP address conflict. The device
with MAC address 000b-824d-cef3 connected to Bridge-Aggregation1 in VLAN 10 and the device with MAC address
000b-824d-cef3 connected to GigabitEthernet1/0/21 in VLAN 10 are using the same IP address 10.10.22.200.
…
Jan 11 05:31:11 CoreSw1 %%10MAC/5/MAC_FLAPPING(l): MAC address 0800-2353-d3e3 in vlan 2368 has flapped from
port Bridge-Aggregation1 to port GigabitEthernet1/0/21 1 time(s).
…
Jan 11 05:31:55 CoreSw1 %%10MSTP/6/MSTP_NOTIFIED_TC(l): Instance 0's port Bridge-Aggregation1 was notified
of a topology change.

 Because of temporary ring occurs, I have multicast traffic peak (attached).

What is the reason of these problems with STP on my switches?

How can I debug and fix this problem?

[sdide]

Hi,

It seems like you run RSTP on all switches, but if you have legacy STP on one of them,  I would look at your topology. As I see it on your drawing, your spanning tree diameter is 10. That could work, but you need to check with the spanning tree timers etc etc. Normally a STP diameter of maximum 7 is recommended. This could explain the behavior you see.

(https://supportforums.cisco.com/discussion/9744871/spanning-tree-maximum-diameter)

If all is RSTP, you might want to see if there is a bad link somewhere (i know you said there weren't any errors on the interfaces but a bad a PHY on a transcieves can cause errors you don't neccesarily see on the electrical side.) .

Regards.

[16again]

You're not allowing vlan1 on trunk, on some STP implementations vlan1 is vital

[SergeyVl87]

Hi!

Today I have the same problem, but there are another interfaces on another switch in syslog messages:

Jan 16 07:54:38 Rack05-PoeSw3 %%10MSTP/5/MSTP_BPDU_RECEIVE_EXPIRY(l): Instance 0's port GigabitEthernet1/0/25
received no BPDU within the rcvdInfoWhile interval. Information of the port aged out.
Jan 16 07:54:39 Rack05-PoeSw3 %%10MSTP/5/MSTP_BPDU_RECEIVE_EXPIRY(l): Instance 0's port GigabitEthernet1/0/26
received no BPDU within the rcvdInfoWhile interval. Information of the port aged out.

[Rack05-PoeSw3]dis stp hist
 --------------- STP slot 1 history trace ---------------
 -------------------  Instance 0   ---------------------
 Port GigabitEthernet1/0/26
   Role change   : Desi->Root
   Time          : 2017/01/16 07:55:30
   Port priority : 0.7848-596b-92e6 3001 32768.d894-0346-5aaa 0
                   32768.d894-0346-5c12 128.25
 Port GigabitEthernet1/0/25
   Role change   : Root->Alte
   Time          : 2017/01/16 07:55:30
   Port priority : 0.7848-596b-92e6 4000 32768.d894-0346-5aaa 0
                   32768.d894-0346-5ac8 128.26
 Port GigabitEthernet1/0/25
   Role change   : Desi->Root
   Time          : 2017/01/16 07:55:30
   Port priority : 0.7848-596b-92e6 4000 32768.d894-0346-5aaa 0
                   32768.d894-0346-5ac8 128.26
 Port GigabitEthernet1/0/26
   Role change   : Root->Desi (Aged)
   Time          : 2017/01/16 07:54:39
   Port priority : 0.7848-596b-92e6 3001 32768.d894-0346-5aaa 0
                   32768.d894-0346-5c12 128.25
 Port GigabitEthernet1/0/25
   Role change   : Alte->Desi (Aged)
   Time          : 2017/01/16 07:54:38
   Port priority : 0.7848-596b-92e6 4000 32768.d894-0346-5aaa 0
                   32768.d894-0346-5ac8 128.26

Because of this, I think, the cause of problem is not in a bad link (cable) or bad SFP transceiver.

Interestingly, the problem occurs usually at night or on weekends, when network traffic is minimal.

As I know, maximum STP disameter = 7 is only recomendation for default STP timers.

I have RSTP configured on all switches with similar STP timers etc.

Some days ago I have changed hello interval (2 sec -> 1 sec) to increase STP stability on network with big diameter.

[Rack05-PoeSw5]dis stp
-------[CIST Global Info][Mode RSTP]-------
CIST Bridge         :32768.d894-0346-5762
Bridge Times        :Hello 1s MaxAge 20s FwDly 15s MaxHop 20
CIST Root/ERPC      :0.7848-596b-92e6 / 2001
CIST RegRoot/IRPC   :32768.d894-0346-5762 / 0
CIST RootPortId     :128.26
BPDU-Protection     :enabled
Bridge Config-
Digest-Snooping     :disabled
TC or TCN received  :195
Time since last TC  :0 days 3h:39m:46s

Is it right: I need to enable "stp root-protection" "stp loop-protection" on all root and alternate ports to prevent loops and increase MaxAge and ForwardDelay timers?

As I know, I don't need to enable vlan1 for RSTP on Comware.

Thanks!

[johnk3r]

Hi !

Below is what is written in the documentation:

HP does not recommend you to manually set the spanning tree timers. Instead, you can specify the network diameter and let spanning tree protocols automatically calculate the timers based on the network diameter. If the network diameter uses the default value, the timers also use their default values.

Configure the timers on the root bridge only, and the timer settings on the root bridge apply to all devices on the entire switched network.

You have performed some settings supression (Multicast / broadcast / unicast)?

stp root-protection must be enabled on the ROOT primary / secondary STP switch

[SergeyVl87]

Hi!

Below is what is written in the documentation:

HP does not recommend you to manually set the spanning tree timers.....

Yes, I know about this, but I can not set diameter = 10

[CoreSw1]stp bridge-diameter ?
  INTEGER<2-7>  Bridge diameter

I have these formulas:

MaxAge = (4*hello + 2*diameter) - 2
ForwardDelay = (4*hello + 3*diameter) / 2

With Hello = 1 and Diameter = 10, Max Age will be 22 and Forward Delay will be 17.

Configure the timers on the root bridge only, and the timer settings on the root bridge apply to all devices

I have changed timers on CoreSw1, but I have not seen changes on another switches:

[CoreSw1]dis stp
-------[CIST Global Info][Mode RSTP]-------
CIST Bridge         :0.7848-596b-92e6
Bridge Times        :Hello 1s MaxAge 22s FwDly 17s MaxHop 20

[Rack05-GigSw1]dis stp
-------[CIST Global Info][Mode RSTP]-------
CIST Bridge         :32768.d894-0369-a443
Bridge Times        :Hello 1s MaxAge 20s FwDly 15s MaxHop 20

You have performed some settings supression (Multicast / broadcast / unicast)?

Only on CoreSw1 and CoreSw2 for some interfaces not used for STP.

stp root-protection must be enabled on the ROOT primary / secondary STP switch

Yes, of cource. I made a typo, I mean "stp loop-protection" on all root and alternate ports to prevent loops.

[Horly]

HI Sergey,

can I check if you are sync'ing all your devices to a NTP server ? 

I am also having an almost exact issue with RSTP on A5500 core and 1910,1920 edge configuration. Again, the issue only surface in the weekends or early morning. 

May I know the tools that you used to capture the broadcast / multi casts traffic ? 

Thanks in advance 

[sdide]

Hi,.

Are there any powersaving settings enabled on the 1910/1920-switches?

Regards

[Mauzr]

Hi, did someone solve this problem ? I have a similar problem with 2x 5130 (IRF stack) as core and 1920 as access. About 2 times per month there is an issue:

log:

1920-access %%10MSTP/5/MSTP_BPDU_RECEIVE_EXPIRY(l): Instance 0's port Bridge-Aggregation1 received no BPDU within the rcvdInfoWhile interval. Information of the port aged out.

5130-core %%10STP/4/STP_DISPUTE: Instance  0's port Bridge-Aggregation23 received an inferior BPDU from a designated port which is in forwarding or learning state.

display stp-history on Access switch:

Port Bridge-Aggregation1
Role change : Desi->Root
Time : 2018/05/31 14:04:38
Port priority : 0.40b9-3caa-73bf 1 32768.40b9-3c65-b6cc 0
4096.40b9-3ca8-6007 128.718

Port Bridge-Aggregation1
Role change : Root->Desi (Aged)
Time : 2018/05/31 14:03:56
Port priority : 0.40b9-3caa-73bf 1 32768.40b9-3c65-b6cc 0
4096.40b9-3ca8-6007 128.718

Thanks and best regards

Mauzr