Lost IP Address for VMKernel iSCSI After Upgrading to ESX 3.5

[Diplomat]

Hi Folks:

I decided to take the plunge and upgrade one of our ESX 3.0.2 servers to ESX 3.5 with a freshly burnt CD. Now I have the following symptom which I need help fixing:

- the IP address seems to have disappeared for the "VMKernel iSCSI" virtual machine port group, i.e. it does not show up in the VI Client (yes, I installed the latest client after the upgrade). Also, I was not able to see the iSCSI volume/lun that the ESX server was seeing before the upgrade. Until that is, I did the following:

- At the service console command line, run esxcfg-swiscsi -e

- At the service console command line, run esxcfg-swiscsi -d

- At the service console command line, run esxcfg-swiscsi -e

- At the service console command line, run esxcfg-swiscsi -s

After this, I could see the volume/lun and was able to browse it and then re-register the VM that I wanted to bring up.

- when doing a "esxcfg-vmknic -l" at the SSH command prompt I do see the following:

Interface Port Group IP Address Netmask Broadcast MAC Address MTU TSO MSS Enabled

VMkernel iSCSI yyy.yyy.yyy.178 255.255.255.192 yyy.yyy.yyy.191 00:50:..... 0 disabled false

vmk0 VMotion yyy.yyy.yyy.177 255.255.255.192 yyy.yyy.yyy.191 00:50:..... 1500 disabled true

- however, the problem now is that there is no way to power up the VM that I re-registered. Everything is greyed out so cannot even select the option to start up the VM. Also, the fact that I cannot see the IP address for the "VMKernel iSCSI" virtual machine port group in the VI Client worries me. I even tried to remove the "VMKernel iSCSI" virtual machine port group but without success - got an error saying it was being used.

Help! Any suggestions welcome.

[RenaudL]

Interface Port Group IP Address Netmask Broadcast MAC Address MTU TSO MSS Enabled

VMkernel iSCSI yyy.yyy.yyy.178 255.255.255.192 yyy.yyy.yyy.191 00:50:..... 0 disabled false

vmk0 VMotion yyy.yyy.yyy.177 255.255.255.192 yyy.yyy.yyy.191 00:50:..... 1500 disabled true

Try enabling your SCSI vmknic:

1. esxcfg-vmknic -e "VMkernel iSCSI"

If this doesn't work, then it's because ESX 3.5 doesn't accept having 2 different vmknics on the same subnet anymore. This caused issues with a lot of people on ESX 3.0.x and we were loose on this rule. We now strictly enforce it.

[Diplomat]

Thanks, but I tried that already and I get the following error:

Error performing operation: SysinfoException: Node (VSI_NODE_net_tcpip_plumb) ; Status(bad0005)= Already exists; Message= Unable to Set

We're not currently using VMotion, so would it make sense to just remove the VMotion port group? What's the simplest way to do that?

Cheers

Steve

[Diplomat]

Actually, can't remove the VMotion port group since it will disable the NFS mount we have to our SnapServer 4500 (which contains the ISO images).

[RenaudL]

Error performing operation: SysinfoException: Node (VSI_NODE_net_tcpip_plumb) ; Status(bad0005)= Already exists; Message= Unable to Set

That's the message you get when you try to create 2 vmknics with the same subnet (I know, it's kind of cryptic).

You have to either use separate subnets or find a way to only use 1 vmknic for all your needs (unfortunately I'm not really familiar with high-level configuration, so I won't be of much help).

[Andyt_bkb]

The same error message exists in Vi3 (ESXi) when adding an address for vmkernel.

in V3.5 & V3.0.1, I can use the same vswitch for vmkernel as the management network (different ip address - same subnet), but inV3i 70348, it complains with that error.

I have tried creating a new vswitch / port group(if thats the same thing), but still gives me the same error.

Currently with V3.0.1 & V3.5 I have:-

1 x nic & vswitch for management network and for vmkernel,

1 x nic & vswitch for virtual machines.

For V3i, does this mean that I have to have:-

1 x nic for management network,

1 x nic for vmkernel network,

1 x nic for virtual machines.

I really hope not !! as that would mean buying extra nics.

please advise.

[jparnell]

I'm also having the same problem.

Ive just started trialling ESXi, and am currently configuring it like our other ESX servers.

When I try to configure a new portgroup for VMotion, it throws the error "SysinfoException: Node (VSI_NODE_net_tcpip_plumb) ; Status(bad0005)= Already exists; Message= Unable to Set"

Traditionaly what we do is have one 'management' vswitch, connected to two uplinks. This 'management' vswitch has two port groups - one for the service console and one for VMotion. The failover order on these portgroups is different, so they use alternate active and standby uplinks.

Ive tried this approach with ESXi, but like i said, it gives me the error above. How should I configure it in ESXi?

James

[Edouard_GAULUE]

What have you done finaly ? Did you get an answer ?

[jparnell]

The only way I can see to do it is to enable Vmotion on the management

port.

Ive tried all sorts - creating a separate vmotion port on a different

portgroup, in a different vswitch, etc, but all result in the error

above.

[hp_Bladefanatic]

Hi Guys!

I have the exact same problem, I can not extend my environment with a vmkernel port for iSCSI, I have alway the error mentioned above, it is only possible to set the IP adress when I use different subnet!

Is there anyone who can explain what happened? What is the reason ? Bug?

thanx

Michael