Although the title of this post is plural, I am only going to discuss one Purple Screen Of Death (PSOD) at this time. I will add more as I run into them and create a separate page if needed.
A few weeks back one of my co-workers (dual CCIE and super awesome guy) came and talked to me about a PSOD he was getting on three of his ESXi hosts (same exact PSOD on each) : “Two File Systems with the same UUID have been detected”
Before I had a chance to look at the issue he came and told me he found the problem; here is what happened.
The three servers in question were Cisco UCS blades. When he received the above PSOD he had configured the blades to boot from SAN and used a separate LUN for each installation. Well, prior to configuring boot from SAN and installing ESXi he had been doing some other testing and installed ESXi on the local drives, but didn’t erase the configurations prior to doing the whole boot from SAN thing.
When you install ESXi the system UUID, which is pulled from the System Management BIOS (SMBIOS), is set as the ESXi host UUID. Using the DCUI console or by vMA/SSH’ing into a host you can use the following command to determine the ESXi host UUID:
Once he wiped the local drive configurations he no longer got the PSOD and ESXi loaded properly. Since he is using Cisco UCS he could have changed the UUID of the host via UCS Manager. There is also a work around posted here by Dave Mishchenko.