WilliamLam.com

Does vCenter Server recycle VM MAC Addresses after Cross vCenter vMotion?

by // 7 Comments

I recently received a question from a customer who was concerned that after a VM has been migrated from one vCenter Server to another using Cross vCenter vMotion, that the original source VM MAC Address could potentially be recycled and re-used at a later point. Back in 2015, I actually wrote about this very topic and the concerns around VM MAC Address duplication after a Cross vCenter vMotion, which I highly encourage folks to check out if you have not seen this article already.

While re-reading the article, I realized that the article had primarily focused on vCenter Servers that were in Linked Mode or under the same vSphere Single Sign-On (SSO) domain and although I did mention the Cross vCenter vMotion across across different vSphere SSO domains scenario, it looks like the details were a tad bit light.

To quickly summarize, when a VM is migrated from a source vCenter Server to the designation vCenter Server, the VM's MAC Address is added to a MAC Address "block list" on the source vCenter Server. This ensures that the VM MAC Address will not be reallocated by the source vCenter Server which would cause a network conflict. This has been the default behavior since vSphere 6.0 and no additional configuration change is required by customers.

[Read more...]

Can you really deploy the vCenter Server Appliance (VCSA) without DNS and NTP?

by // 13 Comments

The simple answer is Yes. Now, you might be wondering why anyone would want to put themselves through the pain without setting up proper DNS and NTP?

Well, not all environments have the luxury of having either outbound connectivity and/or access to basic infrastructure services like DNS and NTP. This may come as a surprise to some, but there are customers out there that need to operate in very unique and constrained environments. One such example of this is typically from customers that need to deploy vSphere in a "dark site" where local infrastructure services like DNS and NTP are not available.


I recently re-validated this deployment model using the latest vSphere 7.0 Update 3 release running on an Intel NUC 11 which had no outbound connectivity and it was only connected to my laptop, which also had no outbound connectivity or access to DNS or NTP. Since this question recently came up from a customer who was looking to automate this, so I ran through the deployment workflow using the VCSA CLI Installer but this should also be possible with VCSA UI Installer as the same options are supported.

OK, so how do you make this work?

[Read more...]

Heads Up - No healthy upstream error with VEBA vSphere UI plugin with vSphere 7.0 Update 3

by // Leave a Comment

After upgrading my homelab to vSphere 7.0 Update 3, I noticed that my VMware Event Broker Appliance (VEBA) vSphere UI Plugin which is included as part of the VEBA Appliance was no longer functioning properly and would display no healthy upstream error message.


I initially thought this might be environmental, since I had just upgraded from lab from vSphere 7.0 Update 2d to 7.0 Update 3. I had reported the issue to our vSphere UI Engineers who built the VEBA UI plugin and while they were looking into the issue, we had another report from a VEBA user who also was hitting the same issue. Today, I got an update from Engineering and it looks like there was a regression in the Envoy service running in the VCSA which had caused this issue. This issue will be fixed in a future patch update for the VCSA, but in the mean time, VEBA users can apply the workaround outlined below.

Note: This workaround is required for vSphere 7.0 Update 3 or later. The issue has been fixed in vSphere 8.0 but if you are running any version of 7.0 Update 3 or newer, you will still need to apply this workaround.

[Read more...]