Diagnosis of bottlenecks in the channel on the same peg?
There are several linux servers on the same peg in the data center, stand by, among them there is one, when he stuck it in another port in the cisk, it turned out, though, that the link on port to Cisco defined as 1Gbit, but the speed of the impact receiving order 200кбит/sec.
"On the other side", sandini convinced that from their side everything is normal, no restrictions and shaping is not necessary and can not be, the more neighboring servers give the stated speed.
What tools you can run a diagnostic channel and to identify in what place plugging?
how to make a diagnosis for network card, assuming that it fails?