Apologies if this is asked often, I did a bit of searching and google-foo but could not find very new news...
I've long been a fan/user of cloudman-managed Galaxy clusters but for the first time I need galaxy in an AWS environment where we absolutely cannot stand up any sort of public facing service or any sort of host with a visible EIC IP address. The hosts can access the internet through NAT etc. but we can't run them in a public VPC or otherwise expose them to the Internet at large.
My experiment last night with cloudman launching inside a private VPC (with NAT'ed internet access and working forward/reverse DNS resolution for all hosts in the subnet) seemed to have failed - the key error in the cloudman logs was an endless iteration of "attempting to learn public hostname" log entries.
I blew away the install when I saw that error message endlessly repeating because there is zero chance our hosts in this VPC subnet will ever have a public FDQN hostname. They all have private IPs and private (but DNS resolvable) hostnames.
Can anyone give a current state update on cloudman inside private VPCs? Is this something I should have debugged further or am I better off just rolling my own AMI and installing galaxy and perhaps the cloudman elements by hand?
Thanks!
-Chris