VMware Cloud Community
insearchof
Expert
Expert
Jump to solution

DWITCH VS VSWITCH

Hello 

 

Not sure if this is the proper Location for the question but here goes.

I am running VMware 6.7 with 8 ESXI Hosts all running 6.7 with Datacenter and Cluster 

I have VM's running Windows 10 Windows 2012 R2  2016 2019 and some Linux machines. All running

Each ESXI Host has 8 network adapters and 7 have network cables 

I setup a Virtual switch for each adapter some Vswitchs have 2 adapters assigned.

A vmkernal is also created for each vswitch.

 

I then decided to start to use Distributed Switches.

I was able to setup them all up with little issue.

Now all my VM's and Hosts are on the Distributed Switches 

All working fine

 

Here is my problem.

 

I have a Web Farm setup with two Nodes and use Windows 2012 R2 NLB Network Load Balance to all my internal users to access the Web Sites.

I have many Web sites on the Web Farm  But Only one Web site is not accessible.

After more research the reason why no one can get to the web site is because we can not ping the ip address of the NLB Cluster.

This started just after I moved all my VM to the Distributed Switch DPG

The NLB Server have two Network Adapters one is the Primary Adapter which is on the Management DSwitch the Second adapter is for the NLB Cluster  and that is on another DSWITCH with a separate subnet The cluster network adapter has two ip addresses assigned to it.  One ip address is on the subnet that the cluster adapters DWITCH is on the second ip address is on the same subnet as the Management DSWITCH.  

This all was working when I had Virtual Switches.

Any ideas or suggestions 

 

Thank you

 

Tom

If this is that the proper board for this let me know so I can change the location

 

Any help would be appreciated 

 

 

 

 

 

 

Reply
0 Kudos
1 Solution

Accepted Solutions
insearchof
Expert
Expert
Jump to solution

Laegre

 

I resolved this issue by moving my Web sites thru my Kemp Load Balancer    no need for the Windows NLB 

 

Thanks 

View solution in original post

7 Replies
scott28tt
VMware Employee
VMware Employee
Jump to solution

Moderator: Moved to vSphere Network Discussions


-------------------------------------------------------------------------------------------------------------------------------------------------------------

Although I am a VMware employee I contribute to VMware Communities voluntarily (ie. not in any official capacity)
VMware Training & Certification blog
Reply
0 Kudos
Lalegre
Virtuoso
Virtuoso
Jump to solution

Hey @insearchof,

I am not an expert in this topic but please take a look at the next KB: https://kb.vmware.com/s/article/1556

There are some recommendations regarding the NLB modes and some configurations that need to be done on the VDS or VSS for them to work. Probably if you migrated, maybe you forgot to configure them.

Reply
0 Kudos
scott28tt
VMware Employee
VMware Employee
Jump to solution

Moderator: Please do not create multiple threads on the same topic.

The second instance you created in vSphere Upgrade & Install has been archived.


-------------------------------------------------------------------------------------------------------------------------------------------------------------

Although I am a VMware employee I contribute to VMware Communities voluntarily (ie. not in any official capacity)
VMware Training & Certification blog
Reply
0 Kudos
insearchof
Expert
Expert
Jump to solution

Lalegre

 

Thanks for the info

On my distributed Switches and Port Groups I can not find this setting

From your DOC

ESXi 6.x, 6.5.X

Please change “Notify Switches” setting to “No” in port groups or vSwitch to prevent RARP packet with vSphere Web Client.

 

My NLB Servers are running Cluster Mode UNICAST

 

 

Reply
0 Kudos
insearchof
Expert
Expert
Jump to solution

Lalegre

 

I found the setting in the port group 

insearchof_0-1604950084027.png

 

 

I changed it from YES to NO 

How long will it take to update? 

 

I tried pinging the nlb cluster address but no response yet

 

 

Reply
0 Kudos
Lalegre
Virtuoso
Virtuoso
Jump to solution

Hey @insearchof,

Please read the KB properly and take your time as it also ask you to set the portgroups with Forged Transmits as Accept. Basically it says this for the ESXi to not compare the source MAC Address that comes from the VM with the one in the Virtual Machine Adapter.

In your case this is important because the NLB cluster masquerade all the source MAC Adress with the Cluster MAC Address and basically if the portgroup is in Reject mode, the ESXi will drop the packet.

Reply
0 Kudos
insearchof
Expert
Expert
Jump to solution

Laegre

 

I resolved this issue by moving my Web sites thru my Kemp Load Balancer    no need for the Windows NLB 

 

Thanks