2016-05-10 04:08 PM
I have a problem whereby we recently started having a flappy nic on one of the 1gb links on the FAS 8020 controllers.
Current setup is 2 controllers each with 4 nics, 2 x 10gb ports and 2 x 1gb ports. Created ifgrp with each of same type. All four nics on the 1gb ports e0c, e0d on both nodes are operating at 100mbps half duplex. Networks have their switches configured as auto negotiate. Only one 1gb nic out of the four is flapping up and down. On the switch side, it shows the nics operating at 1gb full duplex.
From experience i've known misconfiguration of nic interfaces on either the switch or device end to cause problems of flappy nics when configs don't match each end, but what i don't quite understand is why only one nic out the four is behaving this way when they all have the same config. Also why are e0c and e0d on both nodes defaulting to 100mbps half duplex?
1. Are there any logs i can retrieve to trace when this would have started as all this was working fine before.
2. Should i force each of e0c and e0d to full duplex 1gb rather than auto negotiate
I would really like to know what has caused this and when it occured. Your help on this is much appreciated.
Thanks in advance
2016-05-23 10:13 PM
1. AFAIK, i dont think there's any logs u can find on the filer to check the auto-negotiate on nics.
2. by all means, please fix the nics to max speed & full duplex instead of auto-negotiage. we do see many these types of issues in the field.
3. about the auto-negotiation, two sides(switch & node) will try best to use the max speed both supported to negotiate, if not work, they'll try the next lower one. so that's why most of nics are up and running on 100mb half duplex(lowest speed both side support). about the flipping one, i think the negotiation was never succeeded, that's why it's flipping, and as it's flipping, 1gb one switch side should be ignored.