ESXi 5 NIC`s problem

[indexs]

Hello community!

I have installed second network card on the server and ESXi successfully fund it to. One NIC (Build in) is running on DMZ network second NIC (new Installed) is running on local network and is supposed to be used for ESXi management.

So here comes the problem:

After ESXi reboot first NIC is running ok and DMZ virtual server is accessible, but second NIC is not started or something is wrong in other words I can’t ping the local network IP. If i go on site and restart network management on the ESXi server, after that all NIC are working and it is possible to access ESXi management.

Where i can start and do some troubleshooting?

Thank you in advice!

[weinstein5]

A couple of questions - Is the card you added on the HCL? If it is supported, you will be able to create a new virtual switch adn assign the new NIC to that switch ro you can assign to the exisiting switch -

[indexs]

The NIC is added to vSwitch as vmnic1 and it is in use, but u see after reboot it looks like it`s up but not accessable until manual network management restart.

I can look in log files, but where to look?

[a_p_]

Do you use a managed switch for the ESXi host's uplink? In this case please double check the switch port settings and make sure port secuity is not enabled and spanning-tree portfast is configured.

André

[rickardnobel]

indexs wrote:

The NIC is added to vSwitch as vmnic1 and it is in use, but u see after reboot it looks like it`s up but not accessable until manual network management restart.

Could you provide a screenshot of the network setup?

Are both vmnics connected to the same vSwitch? If so, an incorrect physical switch configuration could prevent you from access and when doing the management restart the so called Port ID could change and put the vmkernel on the working physical switch port.

[indexs]

What do you mean by:

"Do you use a managed switch for the ESXi host's uplink?"

I have attached screenshots from ESXi vSvitch0. Maybe i dod`t get it...

[rickardnobel]

This is most likely the issue.

How is this two vmnics connected to physical switches? Same switch or different?

[indexs]

Same switch.

Schreenshot attached

[rickardnobel]

Then there is most likely some difference in the configuration on the two switch ports connected to the host.

Could you access the physical switch configuration? Present a "show run" or similar?

[indexs]

What you mean by:

Rickard Nobel wrote:

Could you access the physical switch configuration? Present a "show run" or similar?

At the moment i`m able connect to ESXi and do what i need to do, but can`t find any "show run"

p.s. just looked at he log files and got this message:

Lost uplink redundancy on virtual switch "vSwitch0". Physical NIC vmnic1 is
down. Affected portgroups:"Management Network", "VM Network".
warning 2012.05.04. 1:04:17 ns.mydomain.lv

followed to this link:

http://kb.vmware.com/selfservice/microsites/search.do?language=en_US&cmd=displayKC&externalId=100954...

This might help?

[rickardnobel]

indexs wrote:

What you mean by:

Rickard Nobel wrote:

Could you access the physical switch configuration? Present a "show run" or similar?

At the moment i`m able connect to ESXi and do what i need to do, but can`t find any "show run"

I mean the physical switch device, outside of your ESXi host. What is the vendor and model?

Could you access this switch through some console cable, telnet, SSH or similar?

[indexs]

Physical switch - Cisco RV042

And yes I can access it by web. It shows connected even when it`s not possible to ping the ESXi after reboot.

[rickardnobel]

Have you done any configuration on the first port on this Cisco device where the ESXi is connected? My guess is that something is similar on the second port.

[indexs]

Actually no configurations where done there it`s now left by default only login, WAN, LAN information and VPN was configuret. I have also turnned off the Firewall till i get the reason why.

After reboot when second NIC is not accessable if disconect first NIC from DMZ and connect it to Cisco LAN i can ping and access ESXi.

I have contacted also Cisco support, but they did not found anything strage. I have runn out any ideas... but still could it be the Cisco fault?

[rickardnobel]

indexs wrote:

Actually no configurations where done there it`s now left by default only login, WAN, LAN information and VPN was configuret.

What is the LAN configuration you have set?

[indexs]

LAN - Static IP starting from 192.168.2.1

Here are no port forwardings, no speed limits, no firewall, no Multiple Subnet

VPN - DHCP IP staring from 192.168.2.201

DMZ Private IP Address : 192.168.2.101

That in configuration...

[rickardnobel]

That is really not much.

So no VLANs, no spanning-tree, port-security or other?

Can you see some port status information? Link state, speed, duplex, dropped packets?

Can you access the physical console of the ESXi host? Could you enable the ESXi Shell and log on locally?

[indexs]

So no VLANs, no spanning-tree, port-security or other?

Nop no nothing.

Can you see some port status information? Link state, speed, duplex, dropped packets?

Yes i can, status shows that cable is connected.

Second question - see attachement.

Can you access the physical console of the ESXi host?

No right now. But i can access it right now usin VPN to Cisco and then usig vSphere client.

Could you enable the ESXi Shell and log on locally?

Yes, Shell access is enabled now.

[rickardnobel]

Can you see some port status information? Link state, speed, duplex, dropped packets?

Yes i can, status shows that cable is connected.

Second question - see attachement.

And the physical ports going into the host are the two that you renamed in the web gui?

Yes, Shell access is enabled now.

Could you enter the ESXi Shell and run "esxtop". Then press "n" to the network view.

Make a note which vmnic the vmk0 is attached to, while working. Try then to restart the management agent (or something else that will trigger the problem) and check with esxtop again if the vmk has moved to the other adapter. That should at least identify which port is not working.

[indexs]

And the physical ports going into the host are the two that you renamed in the web gui?

One NIC (port) is going to LAN the physical server, second NIC is (port) is connected to DMZ, I didn`t renamed anything.

Could you enter the ESXi Shell and run "esxtop". Then press "n" to the network view.

Before and after ESXi Management restart looks the same:

16777222                 vmk0     vmnic1 vSwitch0              1.19    0.00

But i need to be on site to test it more, even reboot the server and then i can see more. If i reboot now i will not be able access ESXi management..