Naming consistency in lag/lacp models

[ashutshkumr]

• In lag_metrics, request column name states oper_states but response column name states oper_status. Need to stick to one name
  • Why can't use up_links instead of member_ports_up ? Since it's more consistent with min_links (and both of these attributes decide up oper_status of lag, hence making them related).
• In lacp_lag_member_metrics,
  • Could we just call it lacp_member_metrics (since LAG is implied from the keyword lacp) ?
  • Could use member_port_name instead of lag_member_port_name
  • Use either in/out or tx/rx/, not both (e.g. lacp_in_pkts vs lacp_rx_errors)
    • IMHO, using tx/rx should be fine: lacp_packets_tx, lacp_packets_rx, lacp_rx_errors
• In set_device_state, member_port_names can be used instead of lag_member_port_names

[ashutshkumr]

@ajbalogh @PrasenjitAdhikary could you please help review wrt to naming consistency ? the idea is to not redundantly use lag and lacp together and make some column names more consistent with what we've already been doing in OTG.

[PrasenjitAdhikary]

@apratimmukherjee to help with reviewing the naming consistency.

[PrasenjitAdhikary]

We had a debate earlier, in the initial proposal document review and decided "lacp member" is incorrect terminology. Instead, the member is applicable for LAG only, and hence use of 'lag' and 'lag-member'

[PrasenjitAdhikary]

In stats name, in/out and tx/rx both are used for consistency with openconfig stats name. Following are the openconfig names:

                    +--ro counters
                       +--ro lacp-in-pkts?               oc-yang:counter64
                       +--ro lacp-out-pkts?              oc-yang:counter64
                       +--ro lacp-rx-errors?             oc-yang:counter64
                       +--ro lacp-tx-errors?             oc-yang:counter64
                       +--ro lacp-unknown-errors?        oc-yang:counter64
                       +--ro lacp-errors?                oc-yang:counter64
                       +--ro lacp-timeout-transitions?   oc-yang:counter64

[ashutshkumr]

We had a debate earlier, in the initial proposal document review and decided "lacp member" is incorrect terminology. Instead, the member is applicable for LAG only, and hence use of 'lag' and 'lag-member'

Is there a reason why we consider lacp member as incorrect terminology ? It seems to have been used in openconfig paths as well (/lacp/*/member). If the terminology is incorrect by the reasoning that lacp is a per-non-aggregated-port protocol, then maybe we should just call the metrics lacp_metrics ? (since in that case member would be redundant)

[PrasenjitAdhikary]

Agreed on the following points, after a discussion: (Prasenjit, Kingshuk, Ashutosh, Subrata, Sourav)

• oper_status : Correct request column name to match with response name.
• member_ports_up : No change.
• lacp_lag_member_metrics : change to “lacp_metrics”
• lag_member_port_name : No change
• in/out/tx/rx : Use tx/rx instead of in/out in OTG model. No change in telemetry name.

[ankur-sheth]

Need this resolved before v1.0.0.