Problems Implementing Redundant Service Console

[gwatson]

Virtual Center is 2.5 build 64912

ESX Hosts 3.0.1

Worked late last night trying to get this to work... but had to give up after

trying lots of solutions. At one point the host lost its warning message about

not having a redundant network management, however VMotion wouldn't work.

So... What I am trying to achieve is having a second Service Console port so

that if someone disconnects the cable (happened before) the heartbeat will

continue to work and it won't turn off the virtual machine... or when the core

is restarted for example the Virtual Machines will continue running.

Implementation went along the lines of installing a physical Cisco switch,

different subnet and VLAN to the original service console, then connecting each

host to this switch. Within Virtual Centre, I created a new Virtual Switch

utilising this NIC port and created as service Console port. I gave it the

correct IP address & subnet mask. It then complained when I tried to change

the default gateway, to this switch, saying there was no VMKernel.. so I also

created this on the same virtual switch. At this point the warning message on

the redundant network management cleared... however VMotion wouldn't work.. HA

did configure though.

I tried lots of other things.... which I can't even remember now.

So.. my question is.. what's the alternative way of doing this or is it down to

a bug. Should I have added this NIC port into the original service console

& vmkernel virtual switch? Should the switch have been on the same subnet

& vlan?

How does VMware deal with two VMkernel ports?

Thanks in advance.

Graeme

[java_cat33]

Can you provide a screenshot of your current ESX network configuration?

[dinny]

Hiya,

Personally I only have one instance of the Service console (i.e. only one port group and IP address on the vswitch) - but have two physical uplinks/NICS connected to it.

One set as active, the other as standby.

These two uplink ports need to be on the same vlan - but ideally coming from separate physical switches (so HA does not kick in if you lose one of the physical switches)

Dinny

[gwatson]

Dinny...

This would give me the same goal I guess.. does it remove the warning message about no redundant network management? I wouldn't of thought so. This is particularly annoying as it can hide other problems..

Graeme

[dinny]

Can't say I recall seeing that message - so I guess it probably does?.

Where is it displayed, and under what circumstances?

Dinny

PS please award points if any answers are helpful or correct

[gwatson]

Are you running version 2.5? It's a 2.5 feature only.

Click on your HA cluster then the Summary Tab. Any configuration warnings? You would also see a warning triangle on your cluster, so fairly obvious.

Graeme

[dinny]

Hiya,

Still on VC 2.02 I'm afraid...

I would still be surprised (and irritated ) if that config did cause such a warning though.

Dinny

[kjb007]

The warning will go away, whether by having multiple service console interfaces on multiple vSwitches, or by having one port with multiple NICs. The multiple part is the key.

-KjB