Wired Intelligent Edge

 View Only
  • 1.  Aruba CX 8325 - Spanning Tree - Starvation

    Posted May 31, 2022 11:25 AM

    Hi guys,

    anyone an idea how to troubleshoot/solve this?

    csw-bu-r02# sh logg -r | i arved
    2022-05-29T09:57:10.414727+02:00 csw-bu-r02 hpe-mstpd[2132]: Event|2008|LOG_INFO|AMM|1/1|CIST starved for a BPDU Rx on port lag53 from 4096:0a000c-010100
    2022-05-21T09:51:39.414683+02:00 csw-bu-r02 hpe-mstpd[2132]: Event|2008|LOG_INFO|AMM|1/1|CIST starved for a BPDU Rx on port lag53 from 4096:0a000c-010100
    2022-05-16T21:58:06.414560+02:00 csw-bu-r02 hpe-mstpd[2132]: Event|2008|LOG_INFO|AMM|1/1|CIST starved for a BPDU Rx on port lag53 from 4096:0a000c-010100
    2022-05-06T11:08:43.414783+02:00 csw-bu-r02 hpe-mstpd[2132]: Event|2008|LOG_INFO|AMM|1/1|CIST starved for a BPDU Rx on port lag53 from 4096:0a000c-010100
    2022-04-17T06:50:45.414692+02:00 csw-bu-r02 hpe-mstpd[2132]: Event|2008|LOG_INFO|AMM|1/1|CIST starved for a BPDU Rx on port lag53 from 4096:0a000c-010100
    2022-04-13T12:07:40.414564+02:00 csw-bu-r02 hpe-mstpd[2132]: Event|2008|LOG_INFO|AMM|1/1|CIST starved for a BPDU Rx on port lag53 from 4096:0a000c-010100
    2022-04-11T02:19:27.414569+02:00 csw-bu-r02 hpe-mstpd[2132]: Event|2008|LOG_INFO|AMM|1/1|CIST starved for a BPDU Rx on port lag53 from 4096:0a000c-010100
    2022-04-10T14:01:24.414679+02:00 csw-bu-r02 hpe-mstpd[2132]: Event|2008|LOG_INFO|AMM|1/1|CIST starved for a BPDU Rx on port lag53 from 4096:0a000c-010100
    2022-04-08T14:33:15.414594+02:00 csw-bu-r02 hpe-mstpd[2132]: Event|2008|LOG_INFO|AMM|1/1|CIST starved for a BPDU Rx on port lag53 from 4096:0a000c-010100
    2022-04-05T20:22:28.414609+02:00 csw-bu-r02 hpe-mstpd[2132]: Event|2008|LOG_INFO|AMM|1/1|CIST starved for a BPDU Rx on port lag53 from 4096:0a000c-010100
    2022-04-05T10:58:41.414564+02:00 csw-bu-r02 hpe-mstpd[2132]: Event|2008|LOG_INFO|AMM|1/1|CIST starved for a BPDU Rx on port lag53 from 4096:0a000c-010100
    2022-04-01T08:44:08.414653+02:00 csw-bu-r02 hpe-mstpd[2132]: Event|2008|LOG_INFO|AMM|1/1|CIST starved for a BPDU Rx on port lag53 from 4096:0a000c-010100
    2022-03-29T05:09:33.414615+02:00 csw-bu-r02 hpe-mstpd[2132]: Event|2008|LOG_INFO|AMM|1/1|CIST starved for a BPDU Rx on port lag53 from 4096:0a000c-010100
    2022-03-27T15:39:52.414685+02:00 csw-bu-r02 hpe-mstpd[2132]: Event|2008|LOG_INFO|AMM|1/1|CIST starved for a BPDU Rx on port lag53 from 4096:0a000c-010100
    

    I did not find any information / cli config for stp starvation/bpdu throttleing on Aruba CX Switches.

    The primary member of Core Switch VSX Pair 2 in Room BU receives the message above.

    It is connected by MCLAG53 with the Core Switch VSX Pair 1 in Room RZ.

    The connection is "dual-homed", so every member is connected with every member on the other side / in the other room.

    All of these links are 40G (Transceiver QSFP+LR4 and singlemode connection with around 300-500m distance).

    The ISL links are 100G, and the links to Distribution Switches are 100G. These Links are using QSFP28DAC1, QSFP28DAC3, and QSFP28SR4 (Multimode MPO/MTP).

    As this new network is still not very used, as is was build up parallel to the old network, the links are not loaded.

    I do not have any idea, why the switch says it is missing the STP BPDU.

    The two VSX Pairs are in the same MSTP Topology and the VSX Pair 1 in Room RZ is the Root Bridge with Priority 4096.

    csw-bu-r02# sh int lag 53
    
    Aggregate lag53 is up
     Admin state is up
     Description : csw-rz
     MAC Address                 : b8:d4:e7:de:f9:00
     Aggregated-interfaces       : 1/1/53 1/1/54
     Aggregation-key             : 53
     Aggregate mode              : active
     Speed                       : 80000 Mb/s
     L3 Counters: Rx Disabled, Tx Disabled
     qos trust dscp
     VLAN Mode: native-tagged
     Native VLAN: 998
     Allowed VLAN List: all
    
     Statistic                          RX                   TX                Total
     ---------------- -------------------- -------------------- --------------------
     Packets                     484985567             58161938            543147505
       Unicast                   103755975             56151612            159907587
       Multicast                 181339372              1699314            183038686
       Broadcast                 199890292               311012            200201304
     Bytes                     73006875428          17447356583          90454232011
     Jumbos                        9511111              2648978             12160089
     Dropped                             0                    0                    0
     Filtered                        27048                    2                27050
     Pause Frames                        0                    0                    0
     Errors                              0                    0                    0
       CRC/FCS                           0                  n/a                    0
       Collision                       n/a                    0                    0
       Runts                             0                  n/a                    0
       Giants                            0                  n/a                    0
    
    csw-bu-r02# sh int 1/1/53
    
    Interface 1/1/53 is up
     Admin state is up
     Link state: up for 2 months (since Thu Mar 10 21:49:10 CET 2022)
     Link transitions: 1
     Description: csw-rz-r08_P_53
     Persona:
     Hardware: Ethernet, MAC Address: b8:d4:e7:de:f9:95
     MTU 9198
     Type QSFP+LR4
     Full-duplex
     qos trust dscp
     Speed 40000 Mb/s
     Auto-negotiation is off
     Flow-control: off
     Error-control: off
     Rate collection interval: 300 seconds
    
     Rate                               RX                   TX        Total (RX+TX)
     ---------------- -------------------- -------------------- --------------------
     Mbits / sec                      0.05                 0.01                 0.06
     KPkts / sec                      0.05                 0.00                 0.05
       Unicast                        0.00                 0.00                 0.00
       Multicast                      0.02                 0.00                 0.02
       Broadcast                      0.03                 0.00                 0.03
     Utilization                      0.00                 0.00                 0.00
    
     Statistic                          RX                   TX                Total
     ---------------- -------------------- -------------------- --------------------
     Packets                     444731259             38878839            483610098
       Unicast                    68160585             37339434            105500019
       Multicast                 179411336              1228391            180639727
       Broadcast                 197159410               311014            197470424
     Bytes                     58978163297          14680671034          73658834331
     Jumbos                        4026723              2535102              6561825
     Dropped                             0                    0                    0
     Filtered                        27032                    1                27033
     Pause Frames                        0                    0                    0
     Errors                              0                    0                    0
       CRC/FCS                           0                  n/a                    0
       Collision                       n/a                    0                    0
       Runts                             0                  n/a                    0
       Giants                            0                  n/a                    0
    
    csw-bu-r02# sh int 1/1/54
    
    Interface 1/1/54 is up
     Admin state is up
     Link state: up for 2 months (since Thu Mar 10 21:49:10 CET 2022)
     Link transitions: 1
     Description: csw-rz-r09_P_54
     Persona:
     Hardware: Ethernet, MAC Address: b8:d4:e7:de:f9:8d
     MTU 9198
     Type QSFP+LR4
     Full-duplex
     qos trust dscp
     Speed 40000 Mb/s
     Auto-negotiation is off
     Flow-control: off
     Error-control: off
     Rate collection interval: 300 seconds
    
     Rate                               RX                   TX        Total (RX+TX)
     ---------------- -------------------- -------------------- --------------------
     Mbits / sec                      0.00                 0.00                 0.00
     KPkts / sec                      0.00                 0.00                 0.00
       Unicast                        0.00                 0.00                 0.00
       Multicast                      0.00                 0.00                 0.00
       Broadcast                      0.00                 0.00                 0.00
     Utilization                      0.00                 0.00                 0.00
    
     Statistic                          RX                   TX                Total
     ---------------- -------------------- -------------------- --------------------
     Packets                      40254565             19283120             59537685
       Unicast                    35595425             18812196             54407621
       Multicast                   1928131               470924              2399055
       Broadcast                   2731009                    0              2731009
     Bytes                     14028740126           2766688451          16795428577
     Jumbos                        5484388               113876              5598264
     Dropped                             0                    0                    0
     Filtered                           16                    1                   17
     Pause Frames                        0                    0                    0
     Errors                              0                    0                    0
       CRC/FCS                           0                  n/a                    0
       Collision                       n/a                    0                    0
       Runts                             0                  n/a                    0
       Giants                            0                  n/a                    0
    


    ------------------------------
    Robert Großmann
    ------------------------------


  • 2.  RE: Aruba CX 8325 - Spanning Tree - Starvation

    Posted Jun 01, 2022 01:35 AM
    Hi Robert

    I couple of questions:
    • which bridge is your root bridge? Is it Core Switch VSX Pair 1? (Prio is 4096; MAC 0a000c-010100)
    • If the above question is a "yes" ... did you check for any VSX inconsistency in the cluster, especially regarding STP config?
    • Can you share the entire config of both VSX clusters, e.g. all 4 switches?
    • what software version are you running?

    Regards,
    Thomas

    ------------------------------
    Thomas Siegenthaler
    ------------------------------



  • 3.  RE: Aruba CX 8325 - Spanning Tree - Starvation

    Posted Jun 01, 2022 03:43 AM

    Hi Thomas,

    Core VSX Pair 1 (csw-rz-r08 and csw-rz-r09) is the STP Root Bridge with priority 4096

    Core VSX Pair 2 (csw-bu-r02 and csw-bu-r03) is the "backup" STP Root Bridge with priority 8192.

    The config on both pairs is consistent as "show vsx config-consistency" and manual comparing by winmerge.

    Even manual comparing of both VSX pairs by winmerge looks good.

    You can find the complete network diagram from a Visio export --> Netzwerk-Redesign.png

    All configs and the output of the "show vsx config-consistency" of both primary members have been attached to this post as Text-Files.

    All Core Switches (CSW) are running GL.10.09.1000 --> Software-Versionen.txt

    The phenomen with STP Starvation can be also seen on the other Aruba CX 8325 and 8360 Switches. At different dates and times.

    It can not be seen on the new client access switches (casw) with Aruba CX 6200F.

    It would be helpful to troubleshoot this behaviour, but also it would be very helpful if is possible to adjust the timers or threshold for STP/BPDU.

    Thanks and kind regards

    Robert



    ------------------------------
    Robert Großmann
    ------------------------------

    Attachment(s)

    txt
    VSX csw-rz.txt   3 KB 1 version
    txt
    Software-Versionen.txt   106 B 1 version
    txt
    csw-bu-r03.txt   14 KB 1 version
    txt
    VSX csw-bu.txt   3 KB 1 version
    txt
    csw-rz-r09.txt   14 KB 1 version
    txt
    csw-bu-r02.txt   15 KB 1 version
    txt
    csw-rz-r08.txt   14 KB 1 version


  • 4.  RE: Aruba CX 8325 - Spanning Tree - Starvation

    Posted Jun 02, 2022 03:56 AM
    Hi, on the drawing it looks like the connection between csw-bu-r02 / csw-bu-r03 and cdsw-bu-r02 / cdsw-bu-r03 should be in blocking state. Ports in blocking should not send BPDUs and csw-bu-r02 / csw-bu-r03 will not get any from them.
     


    ------------------------------
    Arne Opdal
    ------------------------------



  • 5.  RE: Aruba CX 8325 - Spanning Tree - Starvation

    Posted Jun 02, 2022 04:37 AM
    Edited by thomasbnc Jun 02, 2022 04:38 AM
    Hi Robert

    sorry for the delayed response.

    From the configuration I cannot see anything which is obviously wrong. However, from the drawing there is a couple of questions:
    • Did you only write it on the drawing or did you indeed configure multiple MSTP regions throughout your network?
    • It seems as you have multiple, redundant ways from the client access to the core (i.e. distribution layer is connected to the core and among itself). I cannot identify the reasoning behind as in modern network architectures one tries to avoid loops and instead work with MLAGs. This is perfectly feasible in your network. Can you explain a little bit in further detail?

    Given by the loops and STP in the need to block the redundant links and not knowing the running state of the spanning tree this could be a source of problems which result in BPDU stravation.
    Regards,
    Thomas

    ------------------------------
    Thomas Siegenthaler
    ------------------------------



  • 6.  RE: Aruba CX 8325 - Spanning Tree - Starvation

    Posted Jun 02, 2022 05:45 AM

    Hi Arne,

    No csw-bu-r02 and csw-bu-r02 is the Core VSX Pair in Room BU. There is no blocking/STP on the ISL link.

    The red lines (STP blocked) are to the Server and Client Distribution.

    But my main problem/question is the starvation on the csw-bu-r02 coming from the lag53 from the root bridge csw-rz.

    Hi Thomas,

    I have configured a different MSTP Region for every layer (Core, Server, Client and WAN).

    The core knows all VLANs, and the other layers only the relevant VLANs.

    Default Gateway for all VLANs is the primary/active/master Fiirewall Cluster Member (in Room RZ).

    Im using MCLAG wherevever possible, but physically created a loop by cross-connecting the CDSW and SDSW. The main reason is, that I want to avoid Layer2 traffic traversing the core and that traffic between clients will stay in the Client Layer. Same for the Server to Server Traffic between the different rooms.

    If it is possible with aruba at some point, i would like to convert to MPLS. And to get the VLAN Default Gateways from the firewall to the Distribution Layer and switch to Layer3 only on the core.



    ------------------------------
    Robert Großmann
    ------------------------------