某局点S10506X R7596P09设备聚合口二层转发异常

设备型号：S10506X

版本：R7596P09

现场两台S10506X堆叠，二层聚合口666对接对端S10506堆叠设备，终端网关在S10506上，报文在S10506X上做二层转发

现场发现，同网段终端互访不通，关闭两组堆叠核心设备上任一一条链路，只保留一条物理链路的情况下，可以正常互访。

Aggregate Interface: Bridge-Aggregation666 

Creation Mode: Manual 

Aggregation Mode: Dynamic 

Loadsharing Type: Shar 

Management VLANs: None 

System ID: 0x8000, 305f-77e4-b800 

Local: 

 Port Status Priority Index Oper-Key Flag

 FGE1/4/0/6(R) S 32768 1 1 {ACDEF} 

 FGE2/4/0/6 S 32768 58 1 {ACDEF} 

Remote: Actor Priority Index Oper-Key SystemID Flag 

 FGE1/4/0/6 32768 25 15 0x8000, d461-fe05-8e00 {ACDEF} 

 FGE2/4/0/6 32768 26 15 0x8000, d461-fe05-8e00 {ACDEF}

查看物理端口所在单板为EC单板：

  ===============display device verbose=============== 

Slot   Type                  State    Subslot  Soft Ver            Patch Ver

1/0    LSUM1TGS48SH0         Normal   0        S10500X-7596P09     None     

1/1    LSUM1TGS48SH0         Normal   0        S10500X-7596P09     None     

1/2    LSUM2GP48SA0          Normal   0        S10500X-7596P09     None     

1/3    LSUM2GT48SA0          Normal   0        S10500X-7596P09     None     

1/4    LSUM1QGS6EC0          Normal   0        S10500X-7596P09     None

查看聚合666端口底层信息发现底层没有对应端口：

[NW-CORE-probe]debug  port trunk-global c 1 s 4 29a

TG=61

Agg=29a

[NW-CORE-probe]BCM C 1 slot 4 C 0 trunk/show

Device supports 1032 trunk groups:

  1024 front panel trunks (0..1023), 256 ports/trunk

  8 fabric trunks (1024..1031), 16 ports/trunk

trunk 61: (front panel, 1 ports)= dlf=any mc=any ipmc=any psc=portflow (0x9)

[NW-CORE-probe] bcm c 2 s 4 c 0 trunk/show

Device supports 1032 trunk groups:

  1024 front panel trunks (0..1023), 256 ports/trunk

  8 fabric trunks (1024..1031), 16 ports/trunk

trunk 61: (front panel, 1 ports)= dlf=any mc=any ipmc=any psc=portflow (0x9)

查看diagfile发现下发时由于资源不足返回失败：

%@625^Apr 26 20:12:41:158 2022 NW-CORE DRVPLAT/3/DrvDebug: -Chassis=2-Slot=4;

LINE:1722

MOD:TRUNK,TASK:laggd,-FILE id 0xd0000,SLOT:27--:

bcm_trunk_set returns error:No resources for operation, UnitID = 0

原因为EC单板作为堆叠板时，整框重启时堆叠板启动较快，在底层聚合表项下发时，某些资源未准备好，导致下发失败。

R7624P12已经解决这个问题，建议现网升级版本到R7624P12版本。