1

We have a Hyper-V 2012 r2 cluster and I have just upgraded our SCVMM to 2016. The VMM is installed on a VM in the cluster. Our DHCP server is a VM in the cluster with filters enabled. We have configured SCOM to email us whenever someone attempts to connect to our network and we either allow or deny it.

My old VMM server was generation 1. I created a new generation 2 VM for the upgrade. I saw an article where you can migrate from one server to another with different names. I removed SCOM integration, Backed the DB from VMM Console, uninstalled VMM with retain DB option checked and shutdown the VM. After failed attempts at installing on a server with a different name, I reinstalled the operating system on the new VM from scratch and gave it the same name as the old VMM server and started installing VMM. The service should use local system since we are a small business. During the installation, which ran with a domain admin I got an account lockout alert of the user running the installation. Even with the account lockout alert the installation completed successfully (...) Since this happened more than once I checked when the lockout happened and it was during the setspn step of the installation.

My problem is, ever since I installed the VMM server I keep getting DHCP requests. The only reason I say that is because I restarted the server and from the restart it requests DHCP. I even shutdown the server for a day to see what happens and I didn't get any DHCP alerts while the server was down. Every time it is a random address, not vendor specific. And, it sends requests every hour.

Honestly don't know what to do about this...

Rahamim.

  • Hope someone still follows this. I ran Wireshark and found that SCVMM sends a DHCP Discover packet. I tried several ways to disable this but it still happened×¥ – Rahamim Levi Oct 21 '20 at 14:22

1 Answers1

0

Found the culprit!!!

First of all, When installing Hyper-V and using an ISCSI LUNs, the MSiSCSI services is throwing DHCP discover packets when it start. After it finds the NIC that uses the iSCSI targets it uses that NIC and doesn't attempt to search for DHCP until the next reboot \ restart of the service.

Apparently when installing SCVMM server and SCVMM agent several this is what happens: On the SCVMM server the iSCSI initiator service is installed and starts running (like the Hyper-V) which causes the first DHCP discover packets. Than, On all the server that SCVMMAgent service is installed, a DHCP discover packet is broadcasted repeatedly in a constant interval for each server.

The solution: https://download.microsoft.com/download/A/E/9/AE91DEA1-66D9-417C-ADE4-92D824B871AF/uGuide.doc Although this is for server 2008 and 2003, when Creating the MSiSNS Key and adding the DHCPRetries DWORD with the value 0, it causes iSCSI to stop sending DHCP discover packets.

Hope this helps anyone.

Rahamim.