Draft HLD document for portchannel interface #2
Conversation
Satoru-Shinohara
requested review from
kwangsuk,
rathnasabapathyv and
nagarwal03
| # Scope | ||
| - This document describes the high level design of OpenConfig configuration of portchannel interfaces via REST and gNMI in SONiC. | ||
| - This does not cover the SONiC KLISH CLI. | ||
| - This covers only the portchannel interfaces configuration. |
There was a problem hiding this comment.
Don't we support counters?
There was a problem hiding this comment.
We do support counters for interfaces + Ethernet.
However, when comparing E-SONiC vs SONiC-net's DB's on a fresh image, I saw that creating a PortChannel on Community-SONiC Switch did not produce PortChannel in the COUNTERS_DB in a way that was expected.
E-SONiC Showed portchannel inside COUNTERS_PORT_NAME_MAP
"COUNTERS_PORT_NAME_MAP": {
"expireat": 1719361893.0460842,
"ttl": -0.001,
"type": "hash",
"value": {
...
"PortChannel105": "oid:0x2000000000bce"
}
},
While Community-SONiC Showed it inside COUNTERS_LAG_NAME_MAP
"COUNTERS_LAG_NAME_MAP": {
"expireat": 1719362054.8109653,
"ttl": -0.001,
"type": "hash",
"value": {
"": "",
"PortChannel10": "oid:0x2000000000c41"
}
},
Following the OID for both, E-SONiC Produced the proper counters table
"COUNTERS:oid:0x2000000000bce": {
"expireat": 1719361893.0559826,
"ttl": -0.001,
"type": "hash",
"value": {
"PORT_STAT_IF_IN_BITS_PER_SECOND": "0",
"PORT_STAT_IF_IN_OCTETS_PER_SECOND": "0",
"PORT_STAT_IF_IN_PKTS_PER_SECOND": "0",
...
While Community-SONiC Showed No Results
There was a problem hiding this comment.
Once a traffic flows on the given PO interface, the counters should be populated for the OID.
e.g. mock tables to test LAG counters
https://github.com/sonic-net/sonic-utilities/blob/0e6a55ef5eac306ef61d6f0241625a6baee42ab8/tests/mock_tables/asic1/counters_db.json
There was a problem hiding this comment.
But, I don't think you code can handle counters because your callback does not have a translation logic for those attributes from Counter DB to YANG. You can check.
There was a problem hiding this comment.
I just checked that those attributes for LAG counters - SAI_ROUTER_INTERFACE_STAT_IN_ERROR_OCTETS etc. - are not presented in OC yang. So, it's ok.
No description provided.