Open Stack

On Sat, Jan 06, 2018 at 10:21:12AM -0800, Clark Boylan wrote:
> On Sat, Jan 6, 2018, at 10:03 AM, James E. Blair wrote:
> > Hi,
> > 
> > It seems that every time we boot a new server, it either randomly has a
> > hostname of foo, or foo.openstack.org.  And maybe that changes between
> > the first boot and second.
> > 
> > The result of this is that our services which require that they know
> > their hostname (which is a lot, especially the complicated ones) end up
> > randomly working or not.  We waste time repeating the same diagnosis and
> > manual fix each time.
> > 
> > What is the cause of this, and how do we fix this correctly?
> 
> It seems to be an intentional behavior [0] of part of the launch node build process [1]. We could remove the split entirely there and in the hosts and mailnametemplate to use fqdns as hostname to fix it.
> 
> [0] https://git.openstack.org/cgit/openstack-infra/system-config/tree/playbooks/roles/set_hostname/tasks/main.yml#n12
> [1] https://git.openstack.org/cgit/openstack-infra/system-config/tree/launch/launch-node.py#n209
> 
> Clark
> 
We also talked about removing cloud-init, which has been known to modify our
hostnames on reboot.  When I last looked (few months ago) that was the reason
for renames, unsure this time.

I know we also taked about building out own DIBs for control plane servers,
which would move us to glean by default. In the past we discussed using nodepool
to build the images, but didn't want to add passwords for rax into nodepool.o.o.
That would mean a 2nd instance of nodepool, do people think that would work? Or
maybe some sort of periodic job and store credentials in zuul secrets?

PB