Age | Commit message (Collapse) | Author | Files | Lines |
|
In order to unload controller resources place
opendaylight role to a separate node.
Change-Id: I63db4074a9b50d80e8b8b29abd385b37bdb5f9d7
Signed-off-by: Michael Polenchuk <mpolenchuk@mirantis.com>
|
|
* increase the memory/cpu of controller for noha case
* reduce the memory of computes
Change-Id: Iea9351bb846823fae64a662e92f894eca2f6cecb
Signed-off-by: Michael Polenchuk <mpolenchuk@mirantis.com>
|
|
It is possible that some one changes the
config of the vms from a virtual environment.
This is already shown in the elx lab.
To make everyone aware of this we push it
to the default lab as well.
Change-Id: I4e9012b3237838b98321472bb16037aa0aeacfdc
Signed-off-by: Nikolas Hermanns <nikolas.hermanns@ericsson.com>
|
|
This commit does not change the current behavior in OPNFV, the
preconfigured fuel_bootstrap_cli.yaml from OPNFV ISO is still
used to replace the default settings / fuel-menu bootstrap cfg.
The only addition is the possibility to override the preconfigured
fuel_bootstrap_cli.yaml info using DEA.
JIRA: FUEL-155
Change-Id: I4e66b789fdf0a5b1af512a3efc84fedb72ce3b05
Signed-off-by: Alexandru Avadanii <Alexandru.Avadanii@enea.com>
|
|
Currently all scenarios assume interfaces and transformations are the
same for all nodes in the POD, however some PODs may contain nodes that
have different hardware, or where the interfaces are configured
differently.
In this patch we honor the original interfaces and transformations if
they are present in the dea-override.yaml file. The way to add this
information in the dea-override is by having a "nodes:" section with
this information, ie:
nodes:
- id: 1
interfaces: interfaces_1
transformations: transformations_1
- id: 2
interfaces: interfaces_2
transformations: transformations_2
- id: 3
interfaces: interfaces_1
transformations: transformations_1
The node IDs is used to find out this information.
Change-Id: If6ff8ca28b42e043d1bdf91142a4a56ae36e4304
Signed-off-by: Josep Puigdemont <josep.puigdemont@enea.com>
|
|
|
|
Don't make it executable; don't make it world-writable.
Change-Id: I5d117b39e70fff2c1cb5e62d017efe0387b9b5f4
Signed-off-by: Peter Barabas <peter.barabas@ericsson.com>
|
|
Using e1000 driver on fuel master may lead to temporarily loss of
communication [1]. We can use two approaches here:
* use virtio driver
* install e1000-kmod package
virtio driver should be sufficient for our purposes, we don't use
bonding on fuel master.
e1000-kmod package is available on fuel master and provide e1000 kernel
module which does not introduce additional problems. This package is not
installed by default but for this purpose MASTER_NODE_EXTRA_PACKAGES
cmdline option can be set during fuel master installation.
[1]
http://docs.openstack.org/developer/fuel-docs/userdocs/release-notes/fuel_known_issues.html
Change-Id: I3fc1fed5f0b9c780cd9174205a3c31dae92a180e
Signed-off-by: Michal Skalski <mskalski@mirantis.com>
|
|
|
|
ODL Boron use different set of sfc features.
JIRA: FUEL-169
Change-Id: I38c1e518ffbf625a68f879779c01e4abfdb1d672
Signed-off-by: Michal Skalski <mskalski@mirantis.com>
|
|
In order to avoid fetching up non-cluster nodes
filter them out by option of id.
JIRA: FUEL-183
Signed-off-by: Michael Polenchuk <mpolenchuk@mirantis.com>
Change-Id: If0d0a1480d648167f1bcf726f0d6d345d2e00711
|
|
|
|
1. Install CEPH_OSD on 2 compute nodes
2. Delete default 'net_segment_type: tun' definition, for it
is already defined in dea_base.yaml, and it can be updated
by lab dea files
3. Update osd_pool_size from 1 to 2
4. Michal indicates that
https://review.openstack.org/#/c/342717/ set the limitation
: Restrict usage of controller and osd role on single host
Change-Id: I4907482391588fc21ecf60603b1c599413cbc531
Signed-off-by: Julien-zte <zhang.jun3g@zte.com.cn>
|
|
|
|
Both kvm and ovs plugins require specific cpu flags.
Host-model mode will pass most of the host cpu flags to guest.
If a hypervisor is not able to use the exact CPU model, libvirt
automatically falls back to a closest model.
JIRA: FUEL-157
Change-Id: I77ce625e118b6de51cbcdb58e4be5a6dcaecf384
Signed-off-by: Michal Skalski <mskalski@mirantis.com>
|
|
|
|
This reverts commit 34da00c2baa6abc69a812aae4c30a82ebae561bd.
Change-Id: I495fa1f71ac30e259ca63e90ca8b38dd2300d469
Signed-off-by: Michal Skalski <mskalski@mirantis.com>
|
|
|
|
JIRA: FUEL-170
Change-Id: I5abae6fc950cdabe064b5402d9555cedf81cfc23
Signed-off-by: Fedor Zhadaev <fzhadaev@mirantis.com>
|
|
Both kvm and ovs plugins require specific cpu flags.
Host-model mode will pass most of the host cpu flags to guest.
JIRA: FUEL-157
Change-Id: Iff620a4b4f5948e54f6f8db01411d6659f08ec00
Signed-off-by: Michal Skalski <mskalski@mirantis.com>
|
|
Functest framework is trying to access internal endpoints from host
level which currently is not possible in virtual deployments. Move mgmt
network to untagged interface and setup IP address from mgmt subnet on
the linux bridge where mgmt traffic is traversing.
There will be corresponding changes in securedlab repository.
JIRA: FUEL-167
Change-Id: I29b8ebb23a64e39a4e56b27639a87ce2386b9774
Signed-off-by: Michal Skalski <mskalski@mirantis.com>
|
|
* Use fuel-nshovs instead deprecated fuel-nfvovs
* Modified scenario to enable DPDK via Fuel
* Removed hard dependency from specific POD
* Add ability to merge networks lists
JIRA: FUEL-156
JIRA: FUEL-162
Change-Id: Ie609ff140f1de96e54df37dd9b1881afec792317
Depends-On: Ic6072d938e09a40516c32cdde6223d0a21e57403
Signed-off-by: Fedor Zhadaev <fzhadaev@mirantis.com>
|
|
JIRA: FUEL-157
Change-Id: I09e595902b2c467b5522b7b37a772f7ab35c8cea
Signed-off-by: Michal Skalski <mskalski@mirantis.com>
|
|
|
|
|
|
|
|
- Add template file for base DEA
Change-Id: I8055087b992fdc9850852bad9fa07a4e4fde0085
Signed-off-by: Peter Barabas <peter.barabas@ericsson.com>
|
|
Change-Id: I45c42ac65cfbe6562f0035df3375a2231148e22e
Signed-off-by: Peter Barabas <peter.barabas@ericsson.com>
|
|
- fix vertical whitespace
- more consistent function naming
- clean up file writes
- break up some long lines
- typo fixes
- remove duplicate spaces
- unify print's
Change-Id: I5517747ef9f2e39ade7fb553ae2b1547fdf7b9e1
Signed-off-by: Peter Barabas <peter.barabas@ericsson.com>
|
|
JIRA: FUEL-157
Change-Id: I5baafd4395a2d298a24cccda042d5b43824ae7a0
Signed-off-by: Michal Skalski <mskalski@mirantis.com>
|
|
|
|
|
|
- Remove unneeded method
- Write result to a file and not STDOUT
- Add documentation
- Remove trailing whitespace
- Documentation corrections
Change-Id: I7532222d3512380c4f1129bd05dc2ba37b409dc2
Signed-off-by: Peter Barabas <peter.barabas@ericsson.com>
|
|
Fixes https://jira.opnfv.org/browse/FUEL-152
Change-Id: I444bf3aef54ffd53c53431e2795b11b10545f55f
Signed-off-by: Fedor Zhadaev <fzhadaev@mirantis.com>
|
|
Fix typo which result in error while getting default metadata
in case if it's not overridden in scenario.
Fixes https://jira.opnfv.org/browse/FUEL-151
Change-Id: Ibf40f846919155e27da5dc1f778f72afee79ff12
Signed-off-by: Fedor Zhadaev <fzhadaev@mirantis.com>
|
|
We have previously used "pc-1.0" as the qemu-kvm machine definition
for virtual Fuel deploys. On Ubuntu this points to:
pc-1.0 Standard PC (i440FX + PIIX, 1996)
On CentOS 7 this machine definition does not exist, but there's a
match in "pc":
pc RHEL 7.0.0 PC (i440FX + PIIX, 1996) (alias of
pc-i440fx-rhel7.0.0
So, finally, pc is also in Ubuntu:
pc Ubuntu 14.04 PC (i440FX + PIIX, 1996) (alias of
pc-i440fx-trusty)
This change is about exploring whether moving to "pc" will make the
auto deploy work on both platforms with the "pc" machine type.
What still needs to be done in CentOS is to create a symlink from
/usr/libexec/qemu-kvm and /usr/bin/kvm though.
Change-Id: I1cce52a21075950cb59b75709b12d724d759e3e6
Signed-off-by: Stefan K. Berg <stefan.k.berg@ericsson.com>
|
|
|
|
* changes:
deploy: add support for multiple bridges
virtual_fuel: prepare class to allow multiple bridges
|
|
Armband adds similar files, which tend to have long names and only
differ in suffixes (e.g. DEBs for amd64 vs arm64).
Previous change [1] fixed ISO build errors, but failed to address
the deploy.py ISO rebuild (which modifies and recreates the ISO,
calling mkisofs, which complains about clashing file indexes).
This works around ISO rebuild error(s) like:
"genisoimage: Error:
./ubuntu/pool/main/g/golang-gogoprotobuf/golang-gogoprotobuf\
-dev_0.0~git20150828.0.6cab0cc-1~u14.04+mos1_arm64.deb and
./ubuntu/pool/main/g/golang-gogoprotobuf/golang-gogoprotobuf\
-dev_0.0~git20150828.0.6cab0cc-1~u14.04+mos1_amd64.deb
have the same Joliet name"
[1] https://gerrit.opnfv.org/gerrit/#/c/14973/
Change-Id: If58a18ba46343a52ce6e9a0a6a0482c1a3079451
Signed-off-by: Alexandru Avadanii <Alexandru.Avadanii@enea.com>
|
|
Plugin version is required when plugin configuration is overwritten.
Change-Id: I93af002ed5e994c90b9134f8c020a3822e81e42c
Signed-off-by: Michal Skalski <mskalski@mirantis.com>
|
|
Change-Id: I15bb84c35c075037df23746b75b108e2d7b90ede
Signed-off-by: wu.zhihui <wu.zhihui1@zte.com.cn>
|
|
Change-Id: I9db366806ec97b44b0e537ee2d4beb3c8a968b92
Signed-off-by: Nikolas Hermanns <nikolas.hermanns@ericsson.com>
|
|
Change-Id: Ia5a5d008af717f9b2d0e7495bd2180cdad4075e8
Signed-off-by: Michal Skalski <mskalski@mirantis.com>
|
|
SFC scenario requires OpenDaylight Boron.
ODL plugin will be built with both Beryllium and Boron. It will be
possible to choose deployment with Boron instead of default Beryllium.
Change-Id: I1abd22ddca16e710430accd40ad8359be1bd7132
Signed-off-by: Michal Skalski <mskalski@mirantis.com>
|
|
Change-Id: I86cdef9d3e3337d1b16b9154c940f73b0a3d8567
Signed-off-by: Nikolas Hermanns <nikolas.hermanns@ericsson.com>
|
|
The default timeout for SCP is 5 seconds, which is not enough in some
cases.
Change-Id: If44669a96ea4b5ce33441eeea36b8cf66a1cdc4a
Signed-off-by: Peter Barabas <peter.barabas@ericsson.com>
|
|
Give more attempts for zte hardware.
Change-Id: I0906b6fe9e046ddd6436e5246494b89c812a40a2
Signed-off-by: wu.zhihui <wu.zhihui1@zte.com.cn>
|
|
Change-Id: I5c7d9b712c66860bfbe923396e12641c17d439b5
Signed-off-by: wuwb1989 <wuwenbin2@huawei.com>
|
|
deploy.py:
Some Fuel VMs may need more than one network interface. To be able to
provide that, we now allow the user to specify the "-b" paramter
(bridge) multiple times, creating a new NIC for each one of them.
The NICs are created in the same order as they are given in the command
line.
There is no change in behavior from earlier versions, pxebr will still
be the default bridge if none is specified in the command line.
deploy.sh:
To reflect the new capabilities of deploy.py, we introduce the
possibility to specify -B more than once in deploy.sh, and honor that
when calling deploy.py. We also make it possible to specify a comma
separated list of bridges, as in: -B br1,br2.
Change-Id: I1a0100f2cfe755ec6adfeedafb391c2357f46f51
Signed-off-by: Josep Puigdemont <josep.puigdemont@enea.com>
|
|
The VirtualFuel class has now two new methods:
del_vm_nics: Deletes all interfaces from the VM.
add_vm_nic: Adds a NIC to the VM, attached to the specified bridge.
The following method has been deleted:
set_vm_nic: implemented with the two new methods
Apart from the deleted method, no functionality has been changed.
This is just a small but necessary step towards adding support for
configuring more than one NIC in the fuel VM.
Change-Id: I9f02c8163dfb9768510e78d5d5e77a0bb43306fb
Signed-off-by: Josep Puigdemont <josep.puigdemont@enea.com>
|