An administrator cables all interfaces on a newly added AHV host and starts migrating VMs to it. Users report that network connections are much slower to the
VMs after the migration.
What is a possible reason for this issue?
An administrator cables all interfaces on a newly added AHV host and starts migrating VMs to it. Users report that network connections are much slower to the
VMs after the migration.
What is a possible reason for this issue?
The issue of network connections being slower to the VMs after migration is likely due to the host still having the default bond configuration. In a Nutanix AHV environment, the default bond configuration is typically set to active-backup, which provides fault tolerance but not improved network throughput. If the other hosts are configured to use link aggregation, such as active-active bonding, they can utilize all available network interfaces simultaneously, thereby providing higher network throughput. Therefore, the discrepancy in network performance is due to the new host using the slower default bond configuration.
Should be A default bond is active-backup which is the slowest. Chances are the other hosts are configured to use both 10G ports hence why users are seeing faster speeds when connecting to their vm's.
A, correct this. I see so many incorrect answers yet no one has made the needed changes?!
I always go to comments to check what the community is saying :)
Agree its A
yup it should be A
Yep it is A
D : data of the VM migrated are replicated to other nodes (RF2 ou RF3), so network bandwith is used.
default bond is active-backup so will use active ports only.
Definitely A
D doesn't even make sense.
A is Correct
I agree with A. because the default bond is active-backup which is slower.
The correct answer is D, until data resiliency is completed on the CVM (new host CVM), traffic will be passing over the network and not pulling from local disks , yes the default bond is active backup, but in order to saturate a 10GB uplink you would need a crazy amount of traffic so changing to balance tcp, or slb is just increasing the pipe bandwidth, local data will be 20GB pipe any day. I have spoken!
This is incorrect.