Mx5108n port stuck. Broadcom maverick bug?

hello,

We have MX7000 chassis with 740c blades and 4 mx5108n switches. We are running vSphere 6.5u3 on each blade.
Randomly last network interface of any esxi got stucked. (This is the one most heavily used as handling vsan traffic)

At switch and esxi level, port is still  showing up but switch cannot see esxi vmk MAC address.

No error found in os10 standard cli but we found below error in Linux shell in /var/log/syslog:

2020-10-14 00:40:56.292 SW1 base_nas[err]:  [ev_log_t_SAI_SWITCH:Switch Id: 0], NPU: [bcmDPC]_soc_apache_ser_process_ecc: [2674]

2020-10-14 00:40:56.293 SW1 base_nas[err]:  [ev_log_t_SAI_SWITCH:Switch Id: 0], NPU: Unit: 0

PGW BOD CLP0 ECC has multiple ECC errors

2020-10-14 00:40:56.293 SW1 base_nas[err]:  [ev_log_t_SAI_SWITCH:Switch Id: 0], NPU: Unit: 0

PGW BOD CLP0 ECC has multiple ECC errors

2020-10-14 00:40:56.391 SW1 base_nas[err]:  [ev_log_t_SAI_SWITCH:Switch Id: 0], NPU: [bcmDPC]_soc_apache_ser_process_start_err: [3132]

2020-10-14 00:40:56.391 SW1 base_nas[err]:  [ev_log_t_SAI_SWITCH:Switch Id: 0], NPU: [bcmDPC]_soc_apache_ser_process_start_err: [3132]

2020-10-14 00:40:56.392 SW1 base_nas[err]:  [ev_log_t_SAI_SWITCH:Switch Id: 0], NPU: Unit: 0

port 1 start error detected

2020-10-14 00:40:56.392 SW1 base_nas[err]:  [ev_log_t_SAI_SWITCH:Switch Id: 0], NPU: Unit: 0

port 1 start error detected

Once we shut/no shut the interface at switch level, traffic come back. Same done at esxcli level has no effect.

we tried switching traffic to another mx5108n but same problem.

Is it a known bug in OS10.5 ?

Thanks