This kind of regular traffic spike is found on the sfp+ interface (esxi server ->rb5009), but when I change the interface between the server and the router to an electric port, this spike is not found in the data flow
Is it the problem of AOC cable compatibility?
How can I find the source of the problem?
Thanks!
Is it a problem? (or do you see a “problem”) ? I only see a “spike” ?
What type of workload are you running ? Receiving (Rx) 600Mbps non-stop has some challenges if you want to start dissecting that down to the bone…
Videostreams? Do you see effective hickups? complaining users?
It’s a 10G interface, so perhaps something to do with buffering? Your ESXi holds back for a very short period of time and then pushes a bit extra out?
Can you check the NIC on the ESXi side to see if you have the same pattern on the Tx side ?
Using 1.3Gbps（about 120Kpps） data stream test, the peak interval is shorter, which may be related to cache?
Under the same hardware conditions, use 1G electric port instead, and there is no spike （600Mbps）
In addition, is there any method to analyze the data flow sent by Esxi ? （find the source of the spike）
Thank you！
The router was replaced (rb5009_OSv7.3->ccr1036_OSv6.49), and the rest was unchanged. The spike disappeared.
Note that : “Auto negotiation” automatically changed from “not available” to “done”
So, is all this related to the compatibility between AOC cable and router ???
