Hi all,
Having a single ESXi 5 host with 2 VM's on it. Very basic setup.
They're both on the same vSwitch, no VLAN, one physical adapter connected.
I thought VM's on the same vSwitch communicated within that switch and not through my physical network. I see now, that this is not the case.
Here's my setup
192.168.1.1 = Linksys router acts as DHCP/DNS etc.
192.168.1.2 = ESXi mgmt
192.168.1.3 = VM1 OpenIndiana (vmtools installed and working)
192.168.1.4 = VM2 Windows 2008 Server
Copying a file from VM1 to VM2 results in the traffic going through the physical network. How can I change this, so it runs internally?
Thanks for any help
Best regards
Jim
They also need to be on the same portgroup.
Yes, they're on the same portgroup.
Copying a file from VM1 to VM2 results in the traffic going through the physical network
Interesting. This should not be the case, at least not for the copied file.
What you could try to verify that the "copy" traffic stays internally is to temporarily create a second vSwitch with no uplinks and connect the two VM's to this vSwitch.
André
That didn't work either, but then it struck me, I'm running DHCP with static lease on all the servers. When I used a manual IP, it could it fact transfer without the uplink connected.
I've not tried with the uplink connected yet, but I can't imagine that a static IP would be needed for "in vSwitch" transfers??
Best regards
Jim
It should not make any difference whether the IP settings are assigned via DHCP or manually. Maybe there are some options defined on the DHCP server which cause this behavior!? Which options are defined? And which values did you configure for e.g. the subnet mask, the gateway and/or any static routes? Do you see any differences (ipconfig /all and route print) when configured via DHCP and static addressing (except for the lease information)?
What happens if you configure only one of the VM's with a static IP address?
André
Did DHCP configured vlans on the host?
Working now.
I didn't change anything router/dns wise, but I configured a static IP + changed both VM's to use VMXNET3 adapter. I have no idea why or how it's now working as these changes shouldn't matter really, but it does and I'm happy
Thanks for your kind help.
Best regards
Jim
The affected VMs have to use the same type of virtual NIC (i.e.: VMXNET3) as well as be in the same port group on the same vSwitch.
You can also speed it up by enabling VMCI on your ESX host: http://kb.vmware.com/kb/1010806
Oddly, both VM's were using E1000 adapter before I changed them to VMXNET3. Bot were also in the same portgroup, as I only use 1 switch and 1 portgroup.
Thanks I'll give that a go!