You said; 'I would gently suggest this isn't really much of an anecdote. This is like saying "I ran Linux once and it never seemed stable, so I stopped updating

You said; 'I would gently suggest this isn't really much of an anecdote. This is like saying "I ran Linux once and it never seemed stable, so I stopped updating it'. I've been running Linux since 1991. I haven't done so for decades, but I contributed several patches to early Linux development, and I'm quite able to use and maintain Linux in most any circumstances. As the sole administrator of the cluster, I was capable, but not necessarily an expert in all areas. I never had trouble solving the problems that came up, but I was often annoyed that they happened at all. It struck me as an immature product under constant development and refactoring, which often introduced new problems. Through several major releases, lots of important system elements were refactored and deprecated. I should have been more focused on actually using the cluster rather than maintaining it, but it needed a lot of attention. My reference to external DNS means a DNS server not on the LAN, but outside, and only accessible through the WAN. Some of the internal VSphere services and operations are DNS intensive, and cannot tolerate the latency you would get over a 10Mbps WAN connection. I eventually set up a local DNS, which solved most of those problems. When the HPE team came to our site, I worked shoulder-to-shoulder with them for about 20 hours to get the cluster up. At the time, they recommended using a local DNS, so they knew about the issue, but they stopped short of warning me to not go external, probably because they were embarrassed, or more likely because they didn't want liability for any product defects.