On 21/11/13 19:29, Timothy Creswick wrote:
I guess the most important question is what failure mode are you observing? Are the cables effectively DOA? If that's the case then it sounds like the most prudent solution would be to test each cable ourselves before they go into production in future.
Ingress CRC errors at the switchport. Change the cable, problem goes away. It's not uncommon for both server uplinks to report some degree of error (though not always the case).
The balance comes from the time invested, as well as having to over-buy DACs by a factor *just* to keep yourself in working cables. Time and effort that goes towards the (realistic) premium you'll pay for 10GBASE-SR.
Tom
