VMware Cloud Community
markus_herbert
Enthusiast
Enthusiast

Problem with ESX 3.5 host and Console Port

I have 4 esx 3.5 Update 1 servers and 1 VCenter 2.5 also Update 1.

I have lost the connection between the 4 esx hosts and the VCenter Server.

I've seen all 4 esx hosts as disconnected in VCenter (and of course a HA Problem).

Because this is a production system, without vmotion my opportunities to reboot are very limited.

All of the VM's are running and availlable from another interfaces and Ethernet switches.

VCenter 192.168.199.50

esx1 192.168.199.1 and 192.168.199.21

esx2 192.168.199.2 and 192.168.199.22

I opened a console direct on the esx host. What I can do is to ping from 192.168.199.1 to 192.168.199.21.

There are two different interfaces because of redundance to the Console Port.

But ping from 192.168.199.1 to 192.169.199.2 won't work (from one esx1 to esx2) or 192.168.199.50.

I also powered off the Ethernet switch and powered it on - same problem.

The ethernet switch seems to be ok (Link is ok and I also can see data running).

What else I can do (on the direct Console Port) to check what exact the problem is?

Reply
0 Kudos
6 Replies
Frank_Poelert
Contributor
Contributor

Hi Markus,

You setup looks a bit strange to me. Do you really have two IP addresses for each service console?

This might cause the confusion.

For the Service Console you need one virtual switch with two nics (one active, one standby) and only one network connection in the console (vswif0 most likely) and so only IP address.

Please see attachment.

Regards,

Frank

Reply
0 Kudos
weinstein5
Immortal
Immortal

The setup should be ok - by chance have you set vlans up either on your physical switch or on your vswitches? Also are the service console ports on the same vswitch or different?

If you find this or any other answer useful please consider awarding points by marking the answer correct or helpful
Reply
0 Kudos
markus_herbert
Enthusiast
Enthusiast

esx1 192.168.199.1 is the Console Port on Switch 1/NIC 1

esx1 192.168.199.21 is the Secondary Conole Port on Switch 2/NIC 2

Only the VCenter Server has only one NIC (192.168.199.50) and the two switches are connected via a cross over cable.

That worked good for over 2 Month.

I've set no VLAN's.

Reply
0 Kudos
admin
Immortal
Immortal

Have you applied any patches recently?

Are the subnet settings and routing table correct?

esxcfg-vswif -l

route -n

Does the command ifconfig show any packet loss/errors for the vswif interfaces?

Are the servers near each other? Can you connect a cross-over cable and try a ping?

So you cant ping from 192.168.199.1 to 192.169.199.2. Can you ping from 192.169.199.2 to 192.169.199.1 ?

or from 192.168.199.21 to 192.168.199.2 etc?

Are they both connected to the same physical switch? Are their MAC addresses shown in the physical switches CAM table on the correct ports?

If you setup a monitor port on the physical switch on both ports, can you see the pings?

Can do you an arp ping? arping -I vswif0 192....

Message was edited by: appk

Reply
0 Kudos
markus_herbert
Enthusiast
Enthusiast

I found out that one of the switches was crashed and had also crashed the second one.

From the LED's on the frontside everything seems to be ok, but after power off and power on, I can see the desaster.

My esx server worked as expected - leave all vm's powered on.

So my final question:

Is it better to add 2 NIC's also in the VCenter Server to have two complete Ethernet paths to every esx Server?

What I mean

VCenter NIC1 -> Ethernet Switch 1 -> vswif0 (4 x esx-server)

VCenter NIC2 -> Ethernet Switch 2 -> vswif1 (4 x esx-server)

Reply
0 Kudos
Schorschi
Expert
Expert

Dual connectivity to and from the VC to ESX host Service Consoles worth the expense? We decided the lost of the service console NIC is not the end of the world, when it does happen, we just wait for off-hours maintence window, and repair the problem. We have about 600 hosts, for 3 years or so running, and I think we have lost one service console NIC in all that time. Even then, we moved one of our NICs for the VMs bond out of service, and with the dynamic vswitch assignment, and a few CLI commands on the ESX host console, re-established communication with VC server and VMotioned the VMs elsewhere, then fixed the mainboard/replaced, since failed NIC port was embedded.

Reply
0 Kudos