With the advent of high capacity machines (1TB RAM, 4×10 CPU) – consolidation ratio is at its all time high.
Unfortunately, maximum/ limitation of other components would easily be reached.
One good example: we recently experienced VM’s suddenly disconnecting -NICs getting unchecked- from Nexus1000v. We verified there were a lot of free ports left on the 1000v.
After hours of troubleshooting, we finally found the culprit and its stated on the public document:
http://www.cisco.com/c/en/us/td/docs/switches/datacenter/nexus1000/sw/4_2_1_s_v_2_2_1/Resource_Availability/Reference/n1000v_resource_availability_reference.html#wp43988
Nexus 1000v has a limitation of 300 vEthernet interfaces per host.
This means if you have 150 VMs sitting on one host with 2xNICs each- you’ll start to experience the same problem. This number starts to go down the more you add virtual NICs to the VM. 😕
How to verify if you are exhausting this number:
show resource-availability vethports
…so yeah. just a word of advice to anyone using high capacity hosts and running Cisco nexus 1000v.
