06-23-2016 08:31 AM - edited 06-23-2016 10:55 AM
We got Brocade DS-5100 as one of our edge switch sets ISLed to core switch DCX 4S using 2 ports in each fabric. It's been a while that we are regularly getting alerts "Severity latency battleneck detected" for both ISL ports in the DCX. But we have noticed that the botttleneck is always taking place in the same side i.e.Fabric B, while both Fabrics are exactly identical in all ways. Our primary investigations show not much wrong with the ports - both ports are never utilized above 50%. The outcomes of porterrorshow and portstatsshow show the only disc C3 error(specifically- er_tx_c3_timeout) of small counts. How shall we proceed with to resolve this issue? Your any advice and feedback is highly appreciated.
06-23-2016 10:55 AM
Even "small counts" of er_tx_c3_timeout should alert you in my opinion. If you have them on the connected switch as well and there maybe against an end device, the resulting back pressure towards the ISL would explain your severe latency bottleneck messages.