I recently upgraded a vCenter Server instance in a lab from version 7 to 8. In this post I’ll walk through some issues encountered and lessons learned during the process.
Pre-Checks with Skyline Health Diagnostics
In Skyline Health Diagnostics, there is an analysis option for vCenter Upgrade Pre-Check Plugins. This can be found under New Analysis by selecting either product VMware vSphere or VMware vCenter Server. In the current version of Skyline Health Diagnostics (4.0.8 as of this post), this analysis will check a total of 32 possible issues that could impact a vCenter Server upgrade.
In my case, the scan identified a warning named ‘LookupServiceCheck.’ The recommendation was to use lsdoctor to remove some duplicate service registrations. The tools usage is described in this KB article: https://knowledge.broadcom.com/external/article/320837/using-the-lsdoctor-tool.html. The detailed output of Skyline Health Diagnostics also identified specifically which service registrations need to be validated.
After taking a snapshot of the VM, I followed the KB article to remove the duplicate registrations. I then re-ran the vCenter Upgrade Pre-Check Plugins scan, confirmed the issue was resolved, and verified that no new issues were found.
Upgrading vCenter Server
The upgrade from vCenter Server 7.0 to 8.0 was uneventful — which is exactly how I like to describe upgrades. A new vCenter Server Appliance was deployed, the configuration copied over, and logins to the new system, even using active directory users, worked without issue.
It’s difficult to know if the Skyline Health Diagnostics recommendation to run lsdoctor directly contributed to the smooth upgrade. However, the tool was easy to use, and I would definitely include it in my upgrade checklist for future upgrades.
Clearing the arp-cache
Looking in the inventory of the upgraded vCenter Server, I did notice that two test hosts were disconnected. While disconnected test hosts aren’t uncommon, I recalled both hosts being online prior to the upgrade. Interestingly I was able to ping these hosts from my jump box, but not from the vCenter Server appliance.
These hosts were connected to an older consumer grade-switch (TP Link T1600G-28TS) in my lab. I had previously encountered an issue where devices on this switch failed to recognize IP address changes, which I resolved by clearing the arp table. Given that the vCenter Server upgrade workflow involves assigning a temporary IP before reverting to the original, I suspected a similar issue.
I referenced my old notes, which instructed me to SSH into the switch and run a few commands. However, when I attempted to connect using: ssh admin@192.168.10.1, I was met with this error:
Unable to negotiate with 192.168.10.1 port 22: no matching key exchange method found. Their offer: diffie-hellman-group1-sha1A quick search helped me overcome that error. However, I encountered a new error, and then another error, until I found a working connection string:
ssh -oKexAlgorithms=+diffie-hellman-group1-sha1 -oHostKeyAlgorithms=+ssh-dss -c aes256-cbc -m hmac-md5 admin@192.168.10.1Once connected, I ran the following commands:
enable
configure
clear arp-cacheImmediately, the two disconnected hosts came back online. This was a reminder that aging network equipment can introduce unexpected issues. Newer operating systems and SSH implementations have stricter security requirements, and this switch hasn’t kept pace.
Conclusion
The best way to ensure a successful vCenter Server upgrade is preparation. Skyline Health Diagnostics provides an effective way to validate the environment before upgrading and offers guidance on any required remediation steps. Additionally, this experience reinforced the importance of keeping network hardware up to date—technical debt can surface at the most unexpected times, even during routine maintenance.