I was talking to a friend of mine earlier this week about a situation regarding a newly installed VMware vCenter Server Appliance 6.5U1 and was looking for an explanation, which is why this blog post is created. It’s just an information about my findings.
I was setting up a new vCSA 6.5U1 recently and discovered during a reboot of the appliance, that it tried to contact google-hosted time servers which I did not configured at the NTP settings. I was curious why, cause the installation and firstboot was successful and everything seemed to be optimal.
After a reboot I logged in to the vCSA using SSH and looked at
showlog.py messages
During the installation I set up one of my local server to be used as the NTP server. Everything was working fine, but due to the needed restart of the system after joining it to my Active Directory domain, the system was not able to start again.
The log shows local time changed a few times, but did not accept the change and reverted.
I was digging through the vCSA image to find the source of the additional configured time-servers cause I was curious where these settings are coming from. So here is what I found.
If the vCSA did not get a response during boot from the configured NTP server ( in my case it’s caused due temporarily unavailable server-service…) it performs a fallback to the NTP servers configured at the
systemd-timesyncd.conf
which is located at
\etc\systemd\
I enable port 123 traffic from my vCSA to the google-hosted server and the appliance was able to start again. Later I was troubleshooting my local NTP server to be available again 😉
Source: https://wiki.archlinux.org/index.php/systemd-timesyncd
Personal Blog about vAdventures 