@Mats_Hardy thanks for sharing the screenshots, As we understand the VM is connected to the VM Network, with the defined Static IP allocations on the Guest,
For Network isolations/Disconnections: We would highly recommend to follow the Action plan.
1. During the VM Network outage, Do make sure to have the VM console access enabled, and validate if the VM is able to reach to other VMs of the same VM Portgroup (locally) of the same host and other hosts (from outside).
2. Also, see the ARP references on the VM Guest, (# arp -a) and see if the GW IP is getting Populated by default, it shd always be dynamically updated. This validates if the Default Gateway is reachable or not.
If the access to VM is fine, but they are not able to populate the Gateway info, We need to validate the ARP references on the Physical CAM table entries. Do clear the ARP cache, and/or update the same to ensure all the VM MAC references are populated correctly.
Note:
ESXi do not holds any VM ARP Cache details, however it maintains only the ARP cache initiated/associated with the VMKernel instance. ( To check, # esxcli network ip neighbor list)
Refer to the Following articles for more details.
https://kb.vmware.com/s/article/1003893
3. To address any Network disconnections, and packet analysis, we have the following references. As these issues are to be addressed realtime and check from different endpoints ( Network Switch/FI/Firewall/vCenter etc.. )
From vSphere side, we can try to see if the ESXi host which holds the running VM instance, has sent the RARP request to update the MAC table on the Switch. You may simply try to vMotion the VM to another host, to update the entires.
The following actions can be performed at the vNIC level,
a. Disable the Graticious ARP
b. Disable Autoconfigured IP and TCPIP Winsock Resets
c. Added the Local DNS records on the Hosts file, for addressing the DNS and IPv6 issues.
d. Confirmed ipv6 is disabled.
e. DHCP client is also disabled
f. Ensured the VM is running on the latest instance of vmxnet3 Drivers.
For Guest level troubleshooting,
a. We need to address False/Zombie Adapter issues https://kb.vmware.com/s/article/1020718
b. AutoConfiguration IP (DHCP) enabled by default, need to disable them.
# netsh interface ipv4 show inter
# netsh interface ipv4 set interface 'X' dadtransmits=0 store=persistent
c. Registry entries & TCPIP WinSock activities
{refer to the Windows registry settings. "HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\Tcpip\}
Resetting TCP Winsock, followed with reboot.
>netsh i i r r
>netsh winsock reset
Note: Please make a note of the IP settings with G/W address, along with the DNS entries before performing a TCP Winsock Reset.
Regards,
BDV
