do I got you right that most pNICs are directly attached to other pNICs and only a few are attached to a LAN switch?
If my assumption is correct you need to make sure that the portgroups you created only uses that pNIC as active which would allow the system to reach its counterparts, so no standby or multiple active pNICs.
I would also recommend to deactivate auto sense at both ends on those pNICs which are attached to other pNICs.
And of course make sure to use same speed and duplex settings.
Additionally you should also check if some of your equipment use some kind of power savings.
Do the behavior differ between cross connected devices vs those which are attached via your physical LAN switch?
Hope this helps a bit. Greetings from Germany. (CEST)
Troubleshooting loss of connectivity in a multi-VM environment with a vSwitch can be tricky. A few things I would check:
Review the vSwitch configuration and port group settings to make sure nothing was changed accidentally. Verify the correct VLANs and security settings are applied.
Check for any physical network changes on the uplinks, like a trunk port becoming access or vice versa. Also confirm the physical switch configs are correct.
Verify VM network settings like vNIC assignment, IP addressing, default gateway etc. Reconfirm Nothing changed there.
Check for issues like spanning tree blocking redundant uplink ports. Also make sure no upstream network loops were introduced accidentally.
Monitor the vSwitch for errors or packet drops. Tools like esxtop or vCenter can help here.
Consider rolling back the BIOS, firmware or driver updates in case they introduced a regression or compatibility issue.
It's puzzling you lost existing connectivity, so I would thoroughly review configs and logs to find where the change occurred. Let me know if the VMs regain connectivity or if you need any other troubleshooting tips!