diff options
author | Dan Sneddon <dsneddon@redhat.com> | 2017-08-30 11:26:52 -0700 |
---|---|---|
committer | Emilien Macchi <emilien@redhat.com> | 2017-09-01 17:59:24 +0000 |
commit | d88a8dc52ab4331a45a0cd92cf886274cccea73c (patch) | |
tree | 752cf12c4f8f02ec8f595c4f6005156e9495c867 /puppet | |
parent | a3ee4adaee3d59a4cc53daa4b7b5e1d517d4458c (diff) |
Remove ipv6 specific network templates
This change renders the IPv6 versions of the isolated
networks using j2. To allow for backward compatibility,
there will be 2 versions of the network definitions,
<network>.yaml and <network>_v6.yaml. If the ip_subnet
contains an IPv6 address, or if ipv6: true is set on the
network definition in network_data.yaml, then the
<network>.yaml version will contain an IPv6 definition,
otherwise the <network>.yaml will be IPv4, and the
<network>_v6.yaml will be IPv6.
In a future follow-up patch, we will probably only
create the required versions of the networks, either
IPv4, IPv6, not both.
The ipv6_subnet, ipv6_allocation_pools, and ipv6_gateway
settings in the network_data.yaml definition file are
used for the <network>_v6.yaml network definition.
Note that these subnet/cidr/gateway definitions only set
the defaults, which can be overridden with parameters
set in an environment file.
Since the parameters for IP and subnet range are the
same (e.g. InternalApiNetCidr applies to both IPv4/v6),
only one version can be used at a time. If an operator
wishes to use dual-stack IPv4/IPv6, then two different
networks should be created, and both networks can be
applied to a single interface.
Note that the workflow for the operator is the same as
before this change, but a new example template has been
added to environments/network-environment-v6.yaml.
Change-Id: I0e674e4b1e43786717ae6416571dde3a0e11a5cc
Partially-Implements: blueprint composable-networks
Closes-bug: 1714115
(cherry picked from commit dd299f08bd6b1df43760148d83ce9b6e09ba6572)
Diffstat (limited to 'puppet')
0 files changed, 0 insertions, 0 deletions