Known Limitations

When the original flow is egressing out through an aggregated Ethernet (AE) interface, the corresponding sampled sflow frame does not reflect the correct egress port number. This happens only when the flow is egressing out through an AE interface. For non-AE egress interface, this works fine and the sflow frame reflects the correct egress port.PR1647870

ACX7024 ports support 10G/1G/25G multi-rate. When peering with other platform or other vendor devices, For example using SFP-LX10 for 1G connection, the link may remain physically down The reason is Auto-negotiation is not supported in ACX7024 PFE due to vendor limitation. In order to make it work, user has to explicitly configure speed/duplex on both sides, and disable auto-negotiation on the peer side. PR1759804

With T-GM enabled system like ACX7024, ACX7348 and ACX7332, when primary port link comes up late, then PTP system servo state can move from INITIALIZING to PHASE ALIGNED. This does not cause any issue since when in INITIALIZING state, no PTP packets are sent to down streams devices and in PHASE ALIGNED state, PTP packets are sent with all correct clock parameters. When system moves to directly PHASE_ALIGNED, it is assured that the system is internally locked to phase and frequency of GNSS. PR1887028

In ACX7020, DIMM used is Non-ECC Unbuffered Memory. Non-ECC (non-parity) unbuffered memory is a type of computer memory that doesn't have an error-detecting feature and is connected directly to the chipset controller. Due to this EDAC driver is not capable to capture ecc related error packets. but these errors can be detected via mcelog. [vrf:none] root@host-pfe-p1c-06:/var/tmp/mce-inject/mce-inject# cd /var/log [vrf:none] root@host-pfe-p1c-06:/var/log# tail -f mcelog.log Tue Nov 12 23:50:38 PST 2024 antman-pfe-p1c-06 mcelog: {"Socket": { "Status": "3 in 24h", "Description": "corrected Socket memory error count exceeded threshold: 3 in 24h", "Total": 3, "Location": "SOCKET:0 CHANNEL:? DIMM:? []", "Socket": 0, "Corrected_Errors": 3, "Uncorrected_Errors": 0, "Time": 1731484236, "Threshold": "0" } } Due to this feature of DIMM , ECC erros cant be captured by EDAC and same cant be refect in cli root@host-pfe-p1c-06> show chassis memory-ecc-statistics Node Name MBE Countt SBE Count SBE Alarm Th Ce NoInfo Ue NoInfo SBE Insert SBE Log SBE Alarm re0 0 0 10 Disabled Enabled Enabled <==== though alarm is generated - 2024-11-12 23:50:40 PST Minor Memory corrected error on node Re0PR1848612