This was useful when we had clusters in iad2 and rdu3, but it didn't
actually quite work then and it definitely doesn't now. :)
Datacenters like 'aws' do not have their own clusters, they use the one
in rdu3. So, just switch it back to hard coding.
Signed-off-by: Kevin Fenzi <kevin@scrye.com>
This hardware was the 'newest' of the things in iad2.
So, it should have some life left in it.
Some of them are going to be builders/build related things, and others
openqa. We may shuffle this around some more, but this is a first wild
guess.
Signed-off-by: Kevin Fenzi <kevin@scrye.com>
This re-adds a iscsi_client role we had in iad2 back in in rdu3.
When then apply it to bvmhost-p10-01 to login and use a iscsi lun from
the rdu3 netapp. We then move the buildvm-ppc64le vm's to use this iscsi
volume instead of local storage.
As we reinstall those builders they will use the iscsi volume.
Signed-off-by: Kevin Fenzi <kevin@scrye.com>
This should largely have no effect, but makes things easiler to
follow/read. We have 0 rhel7 instances left so removing all those
conditionals.
Also, rename the sshd_config teplate to be .j2 like all our other
templates.
Signed-off-by: Kevin Fenzi <kevin@scrye.com>
This server isn't working and we can't figure out why not. It's
a problem for openQA because we copy the host's DNS config into
'advanced networking' openQA guests, and then when we do a
FreeIPA deployment test, it picks up both DNS servers, tries to
confirm both work, and fails. So we need to take ns03 out until
it's fixed.
Signed-off-by: Adam Williamson <awilliam@redhat.com>
We don't have ipv6 routing setup yet, but are scheduled to work on that
soon. To get ready for that, lets add ipv6 addresses to the (few)
machines that will actually need them.
We do not want to add ipv6 to all hosts. The vast majority of them never
need to talk to the outside world directly and shouldn't have a ipv6
address that can do this.
These few hosts are ones with external nat mappings where it is
desireable that they be able to handle ipv6 connections.
Note that we also do NOT want to add any of these to dns until
they are known working. We also will likely have to adjust nftables
to allow the services on ipv6 that we do on ipv4 (if they make sense).
Signed-off-by: Kevin Fenzi <kevin@scrye.com>
This was done in d97f99c for coreos-ostree-importer. Here I add a
comment with more context to the the coreos-ostree-importer one and
add the same config to the fedora-ostree-pruner.