LTE E-RAB Failures due to MME

Hello Everyone.

I got in hands rather an interesting case related to LTE E-RAB Failures due to MME and I am looking forward to everyone’s insights on this :slight_smile:

Few observations are shown below:

  • E-RAB Failures are exclusively due to MME failures, which can be seen coherent with E-RAB SR

Both sides traces were done as well.

From MME side trace, we did not find abnormal UE CONTEXT RELEASE COMMAND sent by MME which can be mapped with enodeb trace, it is possible that the S1 messages were too much and flow control lost messages.

From eNodeB trace, there is a suspect scenario, when UE initiate the Service request, but at same time receive Downlink data from SGW, the MME default will send UE CONTEXT RELEASE REQ to UE and then paging UE to make UE receive data, it should be controlled by BIT5 of DWORD_EX19 (?)

TX were checked as well, with no major issues to raise.

Check eRAB attempts as well.

You can check max number of DRB configured.

Is it one cell or a cluster?

This issue sounds more related to a capacity issue in the EPC. Do you know if they have enough bandwidth available during peak hours? o Perhaps it could be a capacity issue due to the number of UEs (PDPs during peak hours, for example). These are some of the problems that I have seen when this degradation arises, and it makes sense because it happens during peak traffic hours.

could you please share the enb traces with uu and s1 messages for the affected UE?

for this case, as know most of mme induced failures is from radio side, so I want to know what happens in messages then we could talk about solutions.

have you resolved this ?, i am also seeing same on majority of sites since we migrated swapped the core