I want to ask about the details regarding the network connectvity failure and cluster behaviour in such cases. I am mentioning two such scenarios but other scenarios may also be possible.
If one host in the cluster gets completely disconnected from other nodes at network level but host is up and running what will be the cluster behavior and what will happen to the vms running on that particuar host
If all hosts in the cluster get disconnected from each other at network level but hosts and cvms are up and running, what will happen to the vms running on each host.
Network connectivity failure scenario
This topic has been closed for comments
Already have an account? Login
Login to the community
Login with your account
Enter your username or e-mail address. We'll send you an e-mail with instructions to reset your password.
if you are talking about a host is disconnected from the network "lost network connectivity" so it would lose connectivity to everything "including to other Nutanix Cluster nodes", so in this case the host will be treated as down and the normal Nutanix self-healing process will initiate to re-sync the hosts-disks to restore the Disk resiliency to the configured RF.
in this case all VM's on the hos will be considered as down "from the Hyper-visor level and based on your Hyper-visor Cluster HA configuration the VM's will auto start-up on other available nodes on the clusters.
but if you are talking about loosing the network communications between this node and other Nutanix Cluster nodes but in the same time it is still up and running "and reached over Network" i think this is not a realistic case as all Nuatnix nodes and CVM's are connected on the same Layer2 VLAN.
the same can be applied in Scenario 1 about the CVM being up while the host is down, if the host is down how the CVM will be reachable over Network? the CVM is a VM inside this host and shares the same layer2 VLAN with this Host.
do not forget that Nutanix Cluster will be considered as down if more than the number of applicable to be lost hosts/disks are already lost for this RF "1 for RF2 and 2 for RF3".