I just noticed that while osgeo7 has IP addresses explicitly
configured via netplan, osgeo4 only obtains the IP via DHCP.
Is there any reason for this inconsistency ?
I'm asking because for deploying the machines via ansible
it is useful to be explicit, or we'd need different rules
for dev environment than for production environment.
The only reason is laziness on my part.
When OSUOSL sets up a machine for us, they set it to DHCP out of
convenience.
If the machine has more than one IP, then DHCP doesn't work, so then I reset
it up with netplan as I did with osgeo7 and osgeo3.
I think osgeo4, osgeo8, and osgeo9 are probably all set to DHCP cause they
have only one IP.
I'm fine with you standardizing to always use netplan. I was thinking each
server should probably have at least 2 ips dedicated anyway to simplify
moving things around in the future.
I forget how we do that. I think OSUOSL has a set of ips set aside for us,
though since the containerization effort, a lot of those ips we don't use
anymore.
We should probably find out from OSUOSL if those are still available to us.