jonasbjurel/opnfv-playground

Playground for OPNFV

opnfv-playground's Issues

[CI_PIPELINE] Make the CI-pipeline common for any OPNFV installer

Jump

Jump high

[CI_PIPELINE] Principal port to Fedora-based systems and a precise port to CentOS 7 and RHEL 7.1L

Fix logic for runing toward a certain patch-set

Should preferably not need LF-authentication

Private OPNFV devel. pipeline as a service

When install.sh failed while pip install - deps are never installed

In an env sitting behind a proxy it can happen that pip failed. When running install.sh again pip install will not be called again.

Invalidate cache (-I) seems not to work

[CI_PIPELINE] Verify install.sh on a newly bare installed Ubuntu 14.04 system

Add ci-pipeline clean-up targets

-p Remove virtual resources after ci_pipeline run
-P Remove everyting/restore state to fresh ci_pipeline install status

[CI_PIPELINE] Copying openrc fails

Copying openrc from fuel-master fails resulting in aborted functest stage.

Introduce a way to define deployed OPNFV plugins/options aligned with OPNFV.

Documentation

Update documentation, Wiki and possibly Youtube clips.

[CI_PIPELINE] Clean-up

Add proper clean-up after ci-run:

• Files
• Caches
• KVM VMs

Functest has ceased to work due to refactoring of upstream functest container

The functest container has been integrated with the developers ci-engine and it worked quite neatly the last couple of days, until just recently (today October 20 - or yesterday October 20 ). It seems like the container was updated and the start script was moved to /home/functest/start.sh.
Earlier I used the start script under /home/opnfv/repos/functest/docker/start.sh – but that one no longer has execution rights.
When I try to run functest I get an error

• Error when executing command chmod +x /home/opnfv/repos/functest/testcases/Controllers/ODL/CI/create_venv.sh ……..

What to do?
Would it be possible to version mark the containers – so we get some stability?
It is quite important for us to have the CI pipeline (both the LF as well as our developer one) running through-out our development cycle, and not trying to get it going a week before the release. Ultimately this is an indicator on our progress.

BR/Jonas

Make sure all test results are logged to appropriate logs

Make DEA/DHA selectable on command line

Currently the pipeline uses the static dea.yaml and dha.yaml files (for HA and non HA, respectively). This won't work as we move forward in Fuel versions (for one, wanted_release in dea.yaml will check for a specific version of Fuel), so this needs to be selectable by command line arguments.

One idea is maybe to drop the "-a" argument (Deploy HA config) and instead introduce "-d " argument pointing to a suitable directory containing dea.yaml and dha.yaml, relative to "ci_fuel_opnfv"?

Create a youtube demo

Add needed host vlan to access the Karaf GUI

Automatically adapt install.sh instructions to libvirtd service name

On libvirt-bin/libvirtd: Right now libvirt-bin seems to be the safest bet, but I also seem to recall that libvirtd has been the service name before - let's look into to it, I'm writing an issue for my lousy memory.

Best solution is probably to simply check which service name is the right one in the install.sh logic and present that in the instructions.

Disable iptables traversal of virtual interfaces

When running install.sh, the following entries needs to go into /etc/sysctl.conf (and applied to the running system with sysctl -p):

net.bridge.bridge-nf-call-ip6tables = 0
net.bridge.bridge-nf-call-iptables = 0
net.bridge.bridge-nf-call-arptables = 0 

See http://wiki.libvirt.org/page/Net.bridge-nf-call_and_sysctl.conf for related info.

The above has caused problems when running Fuel 6.1, at least for me.

Be able to consum web links as iso pathes

it would be nice if it is possible to use weblinks, so that the iso is directly downloaded.
./ci_pipeline.sh -a elx_no_ha -i http://artifacts.opnfv.org/fuel/opnfv-2015-12-04_03-04-43.iso

Add propper yardstick infrastructure test-cases

Fix the test pipeline

The test pipeline is not working as intended

R1 Code clean-up

Clean-up, and stream-line the code for the R1 release

Improve ci-status checks

Terminating the CI engine with ^C will sometimes leave a /var/run/fuel/ci-status that is not IDLE. Suggest we add the CI engine PID to that file - if PID not is present, status is definitely IDLE. :)

Order of ci_pipeline.sh arguments wrongly matters

ci_pipeline.sh -a elx_no_ha -b master doesnt work (elx_no_ha seems to be taken as branch)
While ci_pipeline.sh -b master -a elx_no_ha seems to work

Testing of the pipeline for R1 release

[CI_PIPELINE] Ctrl-C is not caught such that post cleaning is triggered

Fix non-ELX dea.yaml for 6.1

[CI_PIPELINE] Create a CI_Pipeline WIKI

[CI_PIPELINE] WWW Front-end radiator

Commit-id shows up wrong

[] Create a main wiki