Cannot login with new install of ESXi 6

[kevanjdm762]

Im having a very frustrating issue with an new ESXi 6 U2 install. The hypervisor installs, I set a password that I know meets the complexity requirements, one PC on the network is originally able to login and manage the machine through the vsphere client. After a short period of time, I can no longer login to the client. Also, from other clients on the network, while I am still connected with the one PC, no others can authenticate. The errors vary from "Cannot login due to an incorrect username and password" which is clearly bs as I can login on the hypervisor console with that username and password and I have one machine that can login at the beginning. The other error is "The request failed because the remote server took too long to respond. The operation has timed out". Then trying again, I usually get the previous incorrect password error. I have another hypervisor running 5.5, on the same switch, same NIC/mgmt config, etc and can connect without any issues what so ever. I've tried the UI, using root and the password I set in the console, again, incorrect password. I've tried changing the password, removing and reinstalling the vsphere client as well as updated from 6.0 to 6.0 U2, all with the same issue. I'm about ready to go back to 5.5 and stop dealing with this. Any suggestions would be greatly appreciated.

[a_p_]

Only a guess. Once the issue occurs, login to the DCUI and restart the management agents.

If this works you may need to check the log files for issues (e.g. service failures, ...).

Are you running ESXi on a supported hardware, or is it a lab/test environment?

André

[unsichtbare]

I would suggest determining if this is an authentication issue or a network issue. It actually sounds like an IP conflict on the LAN to me.

Look in the DCUI, and determine the MAC address of the adapter assigned to the Management Network. Then ping/arp the IP of the ESXi Host in question from many points on the LAN. Don't forget to let the Windows ping run all 4 times before ARP

ping 192.168.100.101

arp -a 192.168.100.101

Then compare the MAC address with the one in the DCUI.

• If there is no response to the ping, then you have a Layer 2 issue
• If the MAC address doesn't match, then you have an IP conflict
• If everything looks right, then consider authentication issues.

[kevanjdm762]

Ok so it is related to the incorrect MAC address for the host, it should be using the maangement interface that has a MAC starting with 00 but it seems to be mapping its IP address to the vm network traffic interface with a MAC of 90. I cant see why this is happening, there is no kernel associated with that interface, no IP assigned, etc. So dont know why this is happening. After each restart, the one machine that usually connect can connect. A ping to the hosts management interface does reply successfully, but the ARP table on each PC on the network shows the incorrect MAC address. Nothing else on the network is using this IP address.

[unsichtbare]

You will have to be much more specific. VMware uses two OUI's:

• 00:50:56
• 00:0C:29

Furthermore, the management VMkernel (usually: vmk0) emulates the MAC address of the physical adapter it is attached to. So you would not expect vmk0 to have either of the VMware OUI's specified above.

IMHO, VMware made a huge mistake when they caused the Management Network to emulate the MAC address of the physical port (sometime around vShere 4.X), as that confuses users as to what a vSwitch/vNIC/pNIC actually are! vmk0 should have a 00:50:56 MAC, just like every other VMkernel NIC you would create!

Anyway, still sounds to me as if you have another device on the same IP address as your ESXi host. Why not try some other IP, after you verify it is available, and see if your problem goes away. Moreover, never regard the ping response "timed out" as meaning there is nothing assigned to that IP, it merely means that a firewall blocked the response.

[kevanjdm762]

When I was referring to MAC address I was referring to the MAC address of the physical adapter that the management interface is on. So for the host that is operating without issue, the vSwitch0 that is just management traffic has a MAC address of 00:25:90:04:28:E6, which corresponds to that actual MAC address of one of the two pNIC's assigned to that vSwitch. On the host not working, the MAC that it should have been showing was 00:25:90:04:44:26, again corresponding to one of the pNIC's assigned to vSwitch0 on the second host but instead was showing 90:E2:BA:97:10:67, which belongs to another pNIC assigned to a completely different vSwitch that is only for VM traffic. The MAC addresses you are talking about, I dont see anywhere on the vSphere client, attached to any pNIC or vNIC. I guarantee there is no other device on this network using that same IP address, no is it configured on any other interface. Changing the IP address from 10.0.0.241 to 10.0.0.242, etc on the management interface, still produces the same effect. When the IP address is changed, all clients that ping that IP address again show the 90:E2 back in the arp table.

[kevanjdm762]

So a reset of the config and reassigning pNIC's has resolved this issue. Maybe a fluke that happened during original network setup, not really sure, but its all playing nice now.

[unsichtbare]

I think I get it.

Any chance you changed-up the physical adapter that the Management Network was assigned to after the host was installed?

In any event, possibly you can use the shell to create a new "Management Network 2," with a temporary IP VMkernel Port on the desired vSwitch - just to regain network management of the host. I might also speculate, you might use the DCUI to change the Management Network NIC, then change it back.

I can get you the shell commands if you like?