Age | Commit message (Collapse) | Author | Files | Lines |
|
We are seeing some instability in latest FDIO packages, so pinning
them for now.
Change-Id: Iaf851bfe2f8c8ba9fe0e2da6979e61f71abeccbe
Signed-off-by: Feng Pan <fpan@redhat.com>
|
|
We don't want the neutorn patch which hardcodes router external
interface to OVS and adds generic NS linux interface driver to be
applied for every scenario.
JIRA: APEX-406
Change-Id: I0fd7ed07a5ff58f9e1565831f4f469c11b501385
Signed-off-by: Tim Rozet <trozet@redhat.com>
|
|
|
|
Change-Id: I16f5defdd47f19095f97414f46cd9e12d3c53b19
Signed-off-by: Tim Rozet <trozet@redhat.com>
|
|
opnfv-tht-pr: 113
opnfv-puppet-tripleo-pr: 16
os-net-config-pr: 5
Change-Id: I0c34889db27fd71e1b73e860850b81a08707c324
Signed-off-by: Feng Pan <fpan@redhat.com>
|
|
We remove and reinstall VPP package for odl_l2-fdio scenario, but
we never removed 80-vpp.conf file during the reinstall, causing
hugepage always being set to 1024.
JIRA: APEX-399
Change-Id: Ifda525517318e4797e4f7809730602ca59b7b6bc
Signed-off-by: Feng Pan <fpan@redhat.com>
|
|
This patch adds support for ODL L3 FDIO scenario.
Introduces a new deploy settings parameter, odl_routing_node, to
indicate the routing node's hostname. This parameter will only
be used for this scenario.
Change-Id: I8df3eb701299007761c6cec42c4ca318b124a0b8
opnfv-tht-pr: 108
os-net-config-pr: 4
Signed-off-by: Feng Pan <fpan@redhat.com>
|
|
This copies the same API jars used for VPP to the Honeycomb to help
workaround the issues we are facing of incompatible Java APIs between HC
and VPP versions. It also locks vpp/honeycomb/odl to known working
packages for L2 scenarios.
Also increases overcloud disk +400MB to accomodate more vpp pkgs.
opnfv-tht-pr: 122
Change-Id: Id803834e1f1fa352778784da9ac246ab9c58a3b1
Signed-off-by: Tim Rozet <trozet@redhat.com>
|
|
Adding the support to run the "Real Time KVM" kernel in the compute
node(s) to reduce the network latency caused by the virtualization of
the network functions. Since the current "Real Time KVM" kernel
doesn't support physical to logical network interface mapping yet, we
need to use the real physical interface name rather than the logical
interface name for all compute interfaces in network_settings.yaml.
Also updates tests for the network settings changes and fixes a bug
where interface was being queried on the controller nics before it
interfaces were validated.
JIRA: APEX-128
Change-Id: Iccb6c9fbdcf7e3b482f4283039ef17c38e0a6df7
Signed-off-by: davidjchou <david.j.chou@intel.com>
Signed-off-by: Tim Rozet <trozet@redhat.com>
Signed-off-by: Dan Radez <dradez@redhat.com>
|
|
Integration work for ONOS in Danube was not completed.
Disabling this scenario for this release.
Disabling completed mostly by commenting out sections enabling
easy re-enabling next release once integration work is completed.
Change-Id: I1d7b457e8234bf6bb299b9908b356040e68370e5
Signed-off-by: Dan Radez <dradez@redhat.com>
|
|
opnfv-tht-pr: 97
opnfv-puppet-tripleo-pr: 10
Change-Id: Idfc109fc984f20b096a23d12864a027146a9c74c
Signed-off-by: Dan Radez <dradez@redhat.com>
|
|
|
|
Builds 6Wind ZRPC, Quagga, and dependencies required for OpenDaylight
with BGPVPN scenario. Packages are built into RPMs, and left on the
overcloud disk in /root/quagga. They are then installed at deploy time
if VPN feature flag is enabled. Note, for proper upstream we should
create a zrpc OOO service, but that can be a follow up patch. For that
we would also need official packaging/build system upstream somewhere.
JIRA: APEX-357
Changes Include:
- Build thrift rpm
- Build c-capnproto rpm
- Build quagga rpm
- Build zrpcd rpm
- Configure bgpd
- Detect deploy with VPN and install pkgs
- Enable zrpcd with systemd
Change-Id: I9825694f46aaab48a3d1cd5fc4d9a24c7370e8fa
Signed-off-by: Tim Rozet <trozet@redhat.com>
Signed-off-by: Nikolas Hermanns <nikolas.hermanns@ericsson.com>
Signed-off-by: Romanos Skiadas <rski@intracom-telecom.com>
|
|
Changes:
- Kernel parameters are now set through first-boot.yaml during deployment. No custom
images will be built. Note that all nodes will be configured the same way, we only
use compute's kernel parameter settings currently (from deploy settings file)
- Add support for VPP interface type in network settings file, it is now possible
to specify vpp_interface type and uio_driver for tenant nic in network settings
file. A new example config, network_settings_vpp.yaml is added.
- Add support for odl_l2-fdio scenario.
Limitations:
- Physical NIC names must be specified in network settings file, numbered nic names
such as nic1 are not supported for fdio scenarios.
- The same kernel parameters will be configured for all nodes. The paramters will be
taken from compute kernel parameter section in deploy settings file.
opnfv-tht-pr: 104
opnfv-puppet-tripleo-pr: 12
os-net-config-pr: 3
Change-Id: Ie9d6151e6e58d11da3c66fbcabe4a0886c3fa152
Signed-off-by: Feng Pan <fpan@redhat.com>
|
|
We used this a while back for additional debug output to prove the quota
setting were correct. Now with composable services, it isn't always
guaranteed that cinder is used. Since we don't use this output for
anything, this patch removes it.
Change-Id: I27ba64761a5ec1bfbe9da0c3e331ba91010d1186
Signed-off-by: Tim Rozet <trozet@redhat.com>
|
|
opnfv-tht-pr: 98
opnfv-puppet-tripleo-pr: 11
Change-Id: I971bb2c94eb9500c81b25adf5019fd02f8e90053
Signed-off-by: Dan Radez <dradez@redhat.com>
|
|
In order to make gluon really deployable some more things
need to be done:
- Add pull request 103 which add yaml files which
point to gluon puppet
- Rename the scenario to odl-gluon similar to bgpvpn
scenario
- Add tht of gluon to overcloud-deploy-functions.sh
JIRA: APEX-379
opnfv-tht-pr: 103
opnfv-puppet-tripleo-pr: 13
Change-Id: I3ccfb5a279f1df21c4da79fc47039b3c0b6e5495
Signed-off-by: Nikolas Hermanns <nikolas.hermanns@ericsson.com>
Signed-off-by: Georg Kunz <georg.kunz@ericsson.com>
|
|
We currently use networking-odl package from opnfv fds repo, this
is no longer needed as networking-odl is packaged in RDO.
Change-Id: Ie9301cd627ab5131d4cd698dd7e4ef435d8e59c0
Signed-off-by: Feng Pan <fpan@redhat.com>
|
|
opnfv-tht-pr: 100
Change-Id: I5db74e9ec74d6d03fde1650739d934ddfeec45ec
Signed-off-by: Nikolas Hermanns <nikolas.hermanns@ericsson.com>
|
|
Command argmument needs two dashes.
Change-Id: I6582dec83f68fe91b0f221ff610aa4149547b32a
JIRA: APEX-296
Signed-off-by: Mark Beierl <mark.beierl@dell.com>
|
|
|
|
|
|
- Also removing a stale file
Change-Id: Ice2f9b4f9961bcd7f05278f11995aa0af237ac8e
Signed-off-by: Dan Radez <dradez@redhat.com>
|
|
Change-Id: I66938516c9b51a94ee398d0ed2696a95c97f992b
Signed-off-by: Tim Rozet <trozet@redhat.com>
|
|
Now in inventory file a user can declare 'disk_device' which will allow
the user to specify which disk on their overcloud server to use for
installation. The variable can be a comma delimited list, which will
search in order post-introspection for the first device on the node that
exists. The default disk used will be sda for deployments.
Currently defining a per node disk is not supported by OSCLI, so
although defined that way in Apex inventory, only the last definition
will be used for all nodes.
Other changes include:
- Various fixes in inventory parsing
- Makes bash writing a common function
- Introspection now enabled for baremetal deployments
JIRA: APEX-296
Change-Id: I330d91eb17408ccfceb7a99c25edbae5ce6d848d
Signed-off-by: Tim Rozet <trozet@redhat.com>
|
|
At present, odl_version is case-sensitive hence opnfv-deploy fail
with 'odl_version: Boron'. So the change fix it.
Change-Id: Ifd611c56b9802b4c5033728e70cce03eb55985cf
Signed-off-by: Tomofumi Hayashi <tohayash@redhat.com>
|
|
Introduces new (hidden) cli option to change composable env config.
This parameter then will use the provided file, for example,
csit-environment.yaml instead of the typical opnfv-environment.yaml.
The effect is using less services required for ODL CSIT snapshots.
Note: this will not work with perf options, but those are not required
currently for ODL CSIT.
CSIT also does not need ceph, so providing a new deploy option to
allow disabling it.
JIRA: APEX-362
Change-Id: If3833b1acceeb23ef9e12c90df6cc9607bd8c724
Signed-off-by: Tim Rozet <trozet@redhat.com>
|
|
JIRA: APEX-346
Change-Id: I9ffa32da906f50441dfc05610d0668e84c992e81
Signed-off-by: Tim Rozet <trozet@redhat.com>
|
|
In newton the heat cli is deprecated:
WARNING (shell) "heat stack-list" is deprecated, please use "openstack stack list" instead
WARNING (shell) "heat resource-list" is deprecated, please use "openstack stack resource list" instead
Change-Id: Ia32f05c72500e5fdd525c32a9b41b2be3cb1599c
Signed-off-by: Dan Radez <dradez@redhat.com>
|
|
|
|
os-odl_l2_nofeature is now deprecated since new netvirt only supports
odl_l3.
JIRA: APEX-299
Change-Id: I3557ecb85fa4cd11bba96aa9410e56a8e8cceb00
Signed-off-by: Tim Rozet <trozet@redhat.com>
|
|
|
|
Also depend on https://github.com/radez/puppet-fdio/pull/14
JIRA: APEX-337
opnfv-tht-pr: 92
Change-Id: Ic3d2f04bfd1983c38e09a2bfa3992a572fc9c6c1
Signed-off-by: Feng Pan <fpan@redhat.com>
|
|
Change-Id: Ie13eaac89c8ce60d7e8c36d21437d536c758688a
Signed-off-by: Feng Pan <fpan@redhat.com>
|
|
JIRA: APEX-334
This regressed when we moved over to installing the replacement ovs
at runtime instead of build time
Change-Id: I0fec6b68be56a3aa41e502c2e09375aca34be22c
Signed-off-by: Michael Chapman <michapma@redhat.com>
|
|
This reverts commit 33bb282dbbea2a2f17c5b3153e60dfff30c4bb15.
Change-Id: I1136c1bd2570a1de50c861c7285434cd244d3b2a
Signed-off-by: Tim Rozet <trozet@redhat.com>
|
|
JIRA: APEX-334
This regressed when we moved over to installing the replacement ovs
at runtime instead of build time
Change-Id: I9624dd32f84e47e3e92dbb091655280b44734e25
Signed-off-by: Michael Chapman <michapma@redhat.com>
|
|
OVS NSH can be installed at deploy time.
Change-Id: Ie875ada99053df95c6534e269f4dd627396489bc
JIRA: APEX-324
Signed-off-by: Dan Radez <dradez@redhat.com>
|
|
OVS with NSH should be installed only if we're enabling SFC in an
ONOS deployment
JIRA: APEX-331
Change-Id: I54826bb2d62b20c42f00c856d8458444f2f795e5
Signed-off-by: Dan Radez <dradez@redhat.com>
|
|
VLAN support allows a set of nodes with a single nic to deploy removing
the need for --flat. flat was not well maintained or tested.
Change-Id: Ieae1c2350c83544ce50048b44300fa0b8e68dd43
Signed-off-by: Dan Radez <dradez@redhat.com>
|
|
- Configure 'netdev' datapath type for ovs agent on compute node
- Set correct permission in OVS
JIRA: APEX-274
opnfv-tht-pr: 83
Change-Id: I94d41375c32f04f4cff872ef85fe8a4514103535
Signed-off-by: Feng Pan <fpan@redhat.com>
|
|
- We now check if vairables (public/private)_(compute/controller)_interface
are set before performing substitution. If those variables are not set,
default values (nic2 for private and nic3 for public) are used.
- Change TenantNIC and PublicNIC variables to be role specific.
opnfv-tht-pr: 80
Change-Id: I4c0c9d5c2239ea1914b5d2de63f2b5c4f63c8fb0
Signed-off-by: Feng Pan <fpan@redhat.com>
|
|
- Update TenantNIC and PublicNIC in numa.yaml
- Use OVS rather than linux bridge for virtual deployment on jump host
- Fix controller and compute IP arrays in T-H-T
- Move networking-vpp repo for vlan rewrite patch
JIRA: APEX-261
opnfv-tht-pr: 79
Change-Id: I4b3b6b09613e7c47869617e15aafec8e7b59d18f
Signed-off-by: Feng Pan <fpan@redhat.com>
|
|
- Add puppet neutron support for configuring both ML2 driver and
networking-vpp agent
- Update THT to invoke new puppet classes.
JIRA: APEX-261
opnfv-tht-pr: 76
Change-Id: Iab21713eb711e9f24fa66cba30c895e43a3e945e
Signed-off-by: Feng Pan <fpan@redhat.com>
|
|
|
|
VPP was overriding kernel settings with its defaults for all deployments
by installing the VPP RPM. This removes that file, and sets the
settings appropriately for the FDIO scenarios only, based on number of
huge pages per profile.
JIRA: APEX-255
Change-Id: If39b220a7c34cbfd673afe1881c1a9bdb3fdc2e0
Signed-off-by: Tim Rozet <trozet@redhat.com>
|
|
L3 Agent needs to use OVS driver for external interfaces and NSDriver
for Internal interfaces.
JIRA: APEX-252
opnfv-tht-pr: 73
Change-Id: Ic96e30421a1f1308c358a644e5602ca738091695
Signed-off-by: Tim Rozet <trozet@redhat.com>
|
|
Previous behavior was detecting the total number of nodes in the
inventory file and defining number of compute nodes as total nodes,
minus 1 or 3 control nodes (depending on HA flag). This patch changes
the behavior to look explicitly for number of compute/control
nodes in the inventory file.
JIRA: APEX-244
Change-Id: I0191bb88e7dc5bebb5f7bfed39661f8047dbd476
Signed-off-by: Tim Rozet <trozet@redhat.com>
|
|
Defaults for virtual deployments was not correct, there were three baremetal
nodes being created by default for a noha deployment.
JIRA: APEX-221
Change-Id: I8b13ffb620d8ebaed4308725edbc7e84279872dd
Signed-off-by: Dan Radez <dradez@redhat.com>
|
|
JIRA: APEX-133
Change-Id: I7d87f2e463921369cd002b768976fd9cc06c8a91
Signed-off-by: Tim Rozet <trozet@redhat.com>
|