We are having an issue in our production environment that some VMs intermittently going out of network. If we migrate (vMotion) to another host the VMs will become active (connected) again.
This issue has been happening since we have updated our esxi hosts to ESXi 6 update 3 build 5224934
But this issue occurs only on one particular cluster. Even though we have updated Esxi hosts in other clusters too.
Same VMs are repeating: loosing network connection again after few days or next day.
As recommended by HPE VMware support, we have updated VMware tools version from 9.. to 10.
Environment Detials:
HP Blade gen8
Vcenter 6
Esxi 6
open-vm-tools
GuestOS : Centos 6
Any help very much appreciated. Thanks
Is it specific VMs or all VMs on this cluster? What are the OS versions on these VMs?
Check if you're using vmxnet3 adapters. There's historically been some issues with E1000.
just start with basics...
- physical connection to blades? any dropped packets on switches/ portchannels /ports etc..
- ESXTOP any dropped packets on pNIC and VM's
- vNIC dropped packets/ crc errors etc..
in the past we had similar issue, but not after upgrades. Just a few VM's dont see each other on different hosts. After days of investigation(opened cases with network and virtualization vendor) and we find out that there is one bad optical cable from chassis to switch. And the rest (3) is good. We were using IP hash as load balancing method.
