How to config the retry count when send a segmented message?

RE: How to config the retry count when send a segmented message?

Mttrinh — Wed, 15 Dec 2021 13:45:29 GMT

Hi,

Provisioning protocol has its own segmentation behaviour, which is independent of transport layer. Does this cause an issue for you? Could you provide a log?

RE: How to config the retry count when send a segmented message?

oplinux — Wed, 15 Dec 2021 02:16:00 GMT

Hi Mttrinh,

Thanks for you prompt response.

We encounter another question. If the provisioner sends several segmented control messages to a node in a short time. eg. Provisioner sends 3 messages in 600ms. the first message don't retry to send out even though there is none ack returned. but the last message will retry to send until get the ack.

a) Is there a message buffer or queue control this process?

b) Why the first message didn't retry to be sent out?

RE: How to config the retry count when send a segmented message?

Mttrinh — Tue, 14 Dec 2021 10:21:35 GMT

Hi,

There is no specific API for this in python. The most correct way to change this would be to use nrf_mesh_opt_set function and NRF_MESH_OPT_TRS_SAR_TX_RETRIES option id (https://github.com/NordicSemiconductor/nRF5-SDK-for-Mesh/blame/master/mesh/core/src/transport.c#L2117-L2123). Even though it says DEPRECATED, there is no specific mesh_opt-function for this case.

The default value could be overwritten here: https://github.com/NordicSemiconductor/nRF5-SDK-for-Mesh/blob/master/mesh/core/include/transport.h#L79