Age | Commit message (Collapse) | Author | Files | Lines |
|
- move cache dir on Fuel Master to /var/cache/opnfv, since /var/lib/opnfv
is deleted during env config;
- allow target image collection on environment deploy failure too;
- create cache directory if missing for bootstrap extraction;
JIRA: ARMBAND-172
Change-Id: Icca650e560913345303ca0430a4d8b38183aa0c9
Signed-off-by: Alexandru Avadanii <Alexandru.Avadanii@enea.com>
|
|
JIRA: ARMBAND-173
Currently only ha scenarios were successfully deployed on virtual pods
The default configuration is for ha scenarios and those have 5 target nodes
while on the noha scenarios there are 4. The noha scenarios failed because
it was trying to parse one extra node.
This patch modifies the dea and dha files for each noha scenario so that the
nodes are properly configured.
Change-Id: I360aeaf8dc92e0ff614367d47d74f6fa61ef2268
Signed-off-by: cristinapauna <cristina.pauna@enea.com>
|
|
Change-Id: I2d30e72cf218251e1b823619501454a65006dd9d
Signed-off-by: Alexandru Avadanii <Alexandru.Avadanii@enea.com>
|
|
JIRA: ARMBAND-172
Change-Id: I9f549e0161d39d8f72fff698cfc26c164b0f10ae
Signed-off-by: Alexandru Avadanii <Alexandru.Avadanii@enea.com>
|
|
Add support for caching deploy artifacts, like bootstraps and
target images, which take a lot of time at each deploy to be built,
considering it requires a cross-debootstrap via qemu-user-static and
binfmt.
For OPNFV CI, the cache will piggy back on the <iso_mount> mechanism,
and be located at:
/iso_mount/opnfv_ci/<branch>/deploy-cache
TODO: Use dea interface adapter in target images fingerprinting.
TODO: remote fingerprinting
TODO: differentiate between bootstraps and targetimages, so we don't
end up trying to use one cache artifact type as the other.
FIXME: This is still a bit hacky in some areas, but we need it ASAP.
JIRA: ARMBAND-172
Change-Id: Iecd293db98c061325643cc61863a48f092bfd491
Signed-off-by: Alexandru Avadanii <Alexandru.Avadanii@enea.com>
|
|
Some leftover chars in https://gerrit.opnfv.org/gerrit/#/c/24901/
were keeping the deploy to work.
Also, the empty folder networks was added. This folder should contain templates
for virtual network config for the targets. For arm we don't need those so we
leave it empty for consistency with other projects and for avoiding changing
the deploy logic
JIRA: ARMBAND-122
Change-Id: Id33e2efacc3ff8d3754c8b29d4bebefd0b385c41
Signed-off-by: cristinapauna <cristina.pauna@enea.com>
|
|
The current fuel arhitecture assumes that all the VMs
(Jenkins slave, Fuel master and Target nodes) are all on
the same server. There is code that allows the Fuel master
to be on a different machine, but nothing for the Target nodes.
In Enea-Arm lab we have a different server for each.
This commit also adds specific templates for a virtual pod in
the Enea-Arm lab.
Therefore, the following changes are done:
- The default libvirt URI is the machine on which the Target nodes are
- The parameter libvirtUri was added in the dha.yaml file for fuel master
- All virsh calls have "-c" option to explicitly specify the uri; this parameter
is set only when libvirtUri is explicitly set in the yaml file
- The scenarios dha override now point to the arm template files
- The network templates are not needed for ARM, as the network infra is provided
by the server on which the Target node VMs reside; the folder is added here
just for consistency
- The target nodes are created with virsh rather than quemu when the nodes are
not on the same machine as Jenkins slave (similar logic to what is done for
remote Fuel master)
JIRA: ARMBAND-122
FIXME: remove hardcodes for
- Network bridges from templates
- Default pool
- libvirtUri from scenario overwrite
Change-Id: Ibf4744e5593f5201550115397024fd8311696074
Signed-off-by: cristinapauna <cristina.pauna@enea.com>
|
|
|
|
JIRA: ARMBAND-169
Change-Id: I4dbcd044c6c146f541e8c005282a4e60af2c3c8a
Signed-off-by: Alexandru Avadanii <Alexandru.Avadanii@enea.com>
|
|
Switch from deploy time building of leveldb + leveldbjni to
using a prebuilt artifact, shipped as a DEB package from Armband
repos.
JIRA: ARMBAND-114
Change-Id: Ic7a52c78e803c1a6ebda1f127774959a0880a5a3
Signed-off-by: Alexandru Avadanii <Alexandru.Avadanii@enea.com>
|
|
Packetary solves "libasound2" with "liboss4-salsa-asound2" from universe,
leading to openjdk failing to install.
Work around this by altering recently introduced packetary-driven local
mirror build to include OPNFV plugin deps into the "main" package list.
JIRA: ARMBAND-167
JIRA: ARMBAND-35
Change-Id: I040a7117c7ae809933a400ae00937ab8fcc1ea0e
Signed-off-by: Alexandru Avadanii <Alexandru.Avadanii@enea.com>
|
|
We need to explicitily add our Armband leveldb patching prereq
tools, like maven, git, make etc. to the ODL plugin's packages.yaml,
so they are included in the offline Ubuntu mirror.
While at it, also add openjdk-8-jre amos version additional dep:
- libasound2
JIRA: ARMBAND-167
Change-Id: I17fa7a4727b8a553f5a46ab13267380a8864c728
Signed-off-by: Alexandru Avadanii <Alexandru.Avadanii@enea.com>
|
|
During the previous attempt to fix the local mirror rebuild by
reworking `atp-ftparchive` configuration, we removed the arch
list from the main configuration, which led to the local Release
file missing "Architectures" entries.
This led to "/etc/fuel_openstack_arch" not being populated,
binfmt not starting on the first boot, and bootstrap build failing.
Bring back the "Architecture" config line to restore the old
behavior.
JIRA: ARMBAND-163
Change-Id: Ica64fcd60ed4e0185d2cab5457603b42d3d55af8
Signed-off-by: Alexandru Avadanii <Alexandru.Avadanii@enea.com>
|
|
|
|
NOTE: Temporarily patch "packetary" in Armband until we upstream
two small fixes.
Bootstrap build, deploy target image build:
- Switch from public Ubuntu mirrors to 10.20.0.2.
JIRA: ARMBAND-35
Change-Id: I14645cfeebb486063e7b8198de273f68285b11cd
Signed-off-by: Alexandru Avadanii <Alexandru.Avadanii@enea.com>
|
|
JIRA: ARMBAND-96
Change-Id: I28ab3030d3f31c5734508838985412a33334570a
Signed-off-by: Alexandru Avadanii <Alexandru.Avadanii@enea.com>
|
|
JIRA: ARMBAND-163
Change-Id: I3fd5a8a826372afc323665abe229c7e67991e31a
Signed-off-by: Alexandru Avadanii <Alexandru.Avadanii@enea.com>
|
|
Slightly edit patch context to match upstream changes [1, 2].
[1] https://gerrit.opnfv.org/gerrit/#/c/22559/
[2] https://gerrit.opnfv.org/gerrit/#/c/21823/
Change-Id: Ibdb3b557f3e482f74d345a1e0e1eedda26879179
Signed-off-by: Alexandru Avadanii <Alexandru.Avadanii@enea.com>
|
|
Armband does not yet have DEB/RPM repo snapshots, so revert
upstream Fuel@OPNFV change enabling them.
JIRA: ARMBAND-155
Change-Id: Ib08c36bf28a66ef38d8b33031dcbf363fb34123c
Signed-off-by: Alexandru Avadanii <Alexandru.Avadanii@enea.com>
|
|
|
|
Now that [1] has been merged, we should stick to upstream behavior.
[1] https://gerrit.opnfv.org/gerrit/#/c/22975/
This reverts commit f653f97dd2591f4cd5bcb4e76eb2134a93f6d460.
Change-Id: I5f23dd6e1b113e8a7d06995e479287ff196410dd
|
|
While at it, also upstream:
"build: Use OPNFV_GIT_SHA for ISO preparer ID"
JIRA: ARMBAND-146
Change-Id: Ifcd7b3041d172aa492e26477c0872cd96efd90ba
Signed-off-by: Alexandru Avadanii <Alexandru.Avadanii@enea.com>
|
|
Fuel Master VM template has been updated to use:
- 4 VCPUs (was 8 VPCUs);
Related-bug: ARMBAND-98
Change-Id: I1a587f0efa06850885ff8cf938ebb9775af8f7a8
Signed-off-by: Alexandru Avadanii <Alexandru.Avadanii@enea.com>
|
|
We recently removed blacklists for a few mirrors, on account
of upstream Fuel@OPNFV fixing mirror selection in [1].
Although the mechanism will be polished a little, it works very
nicely; pointing us to [2], where we seem to trigger some flood
protection at sync, leading to dropped connections and failed builds.
Keep using the official Ubuntu Archive mirror [3] until we sort this out.
Fixes: ARMBAND-136
[1] https://gerrit.opnfv.org/gerrit/#/c/22195
[2] http://mirrors.se.eu.kernel.org/ubuntu/
[3] http://archive.ubuntu.com/ubuntu/
Change-Id: If9247f829ee314d4c71dafa84b6cd13b384484f0
Signed-off-by: Alexandru Avadanii <Alexandru.Avadanii@enea.com>
|
|
While refactoring the patching mechanism, take care of:
- Sync submodule handling with Fuel@OPFNV;
- build: Investigate/prepare for moving patches to Fuel@OPNFV;
- build: Investigate divergent fuel-mirror;
- ISO build: cacheid for Fuel comps should not depend on
Armband git commit;
CHANGE:
Rename/shuffle patches while grouping them in "features",
preparing for upstreaming them to Fuel@OPNFV and beyond.
CHANGE:
Allow linking patches for better representing the dependency
between one patch and different features.
e.g. 0001-Add-arch-to-nailgun-release-and-target-image.patch:
- part of `multiarch-fuel`, because it extends Fuel;
- part of `direct-kernel-boot`, as arch is required for that;
- part of `cross-bootstrap`, target image is arch-dependent;
NOTE: Patch links are not staged to Fuel@OPNFV, they only serve
as markers that a specific patch is part of a feature.
CHANGE:
Kill all Fuel component submodules, now handled in Fuel@OPNFV:
- fuel-agent
- fuel-astute
- fuel-library
- fuel-mirror
- fuel-nailgun-agent
- fuel-web
CHANGE:
Move armband-fuel-config.mk to armband git root.
FIXME: m1.micro-Increase-profile-RAM-size-to-128MB.patch is NOT
part of `cross-bootstrap` feature, but patch context says so ...
FIXME: 0001-Add-arm64-deb-repositories-setup.patch is broken at
`make patches-export` by removing spaces at EOL.
v2 -> v3:
* Phony patch support (links to show a patch belongs to a feature);
* Updated README.md
v3 -> v7:
* Re-export Fuel submodules & plugins patches (update patch context);
* Update Cavium mail addresses (s/caviumnetworks.com/cavium.com/);
* Ignore submodule changes;
* Add armband git repo info to gitinfo_fuel.txt at build time;
Implements: ARMBAND-136
Closes-bug: ARMBAND-95
Closes-bug: ARMBAND-93
Closes-bug: ARMBAND-92
Change-Id: I1a236d9f43b2e6dca22055911f696b43c22b5973
Signed-off-by: Alexandru Avadanii <Alexandru.Avadanii@enea.com>
|
|
Since we are now building inside docker containers, it is no longer
necessary to use `python-debian` from pip repo instead of Ubuntu.
Also, since Fuel@OPNFV improved our mirror selection in [1], we
can drop our custom blacklist entries for Ubuntu SE mirrors, as
well as repeating the repo mirror build in case of failure.
[1] https://gerrit.opnfv.org/gerrit/#/c/22195
Change-Id: I62890b6ae7977cd8a39c06ce17c8964e6cb1dddb
Signed-off-by: Alexandru Avadanii <Alexandru.Avadanii@enea.com>
|
|
Change-Id: If6a3472a2f8fe64db2469536fe4d12d88e7a61f0
Signed-off-by: Alexandru Avadanii <Alexandru.Avadanii@enea.com>
|
|
This reverts commit acc93911c7d1d4be5ba5cd04e731626d01d59598.
Change-Id: I001cd72506ec99ae2c41bd30d3dd36dd3de06217
|
|
Fuel Master VM template has been updated to use:
- 4 GB RAM (was 8 GB RAM);
- 2 VCPUs (was 8 VPCUs);
Related-bug: ARMBAND-98
Change-Id: Ib95e87e8ed0035b2a7e69a95f67a4046d724d624
Signed-off-by: Alexandru Avadanii <Alexandru.Avadanii@enea.com>
|
|
Change-Id: I404791b14603b4ebdd0be1010412ec77cf75a478
Signed-off-by: Alexandru Avadanii <Alexandru.Avadanii@enea.com>
|
|
OVS-DPDK-NSH is not supported on arm64 for now, see [1]
Closes: ARMBAND-46
[1] https://jira.opnfv.org/browse/ARMBAND-86
Change-Id: I79b33bab7348e297c06ff4bb2bf7cc23f21eb7e3
Signed-off-by: Alexandru Avadanii <Alexandru.Avadanii@enea.com>
|
|
On EFI systems deploys, altering the boot order via IPMI causes
issues on some specific targets. Until we agree on an uniform
solution for EFI systems + IPMI boot order handling, just skip
altering it via IPMI.
Related-bug: ARMBAND-71
Change-Id: I7790e1e8e3540d5ab10b8d7165316a04cf58faa4
Signed-off-by: Alexandru Avadanii <Alexandru.Avadanii@enea.com>
|
|
Change-Id: I1281b0b6b9d631e371484795a53113032ebbc83c
Signed-off-by: Alexandru Avadanii <Alexandru.Avadanii@enea.com>
|
|
SE mirrors are acting up lately, blacklist them:
- ftp.availo.se
- ftp.lysator.liu.se
Change-Id: If4cc7c207fa7ce39821cc8a8f754a48f573ff374
Signed-off-by: Alexandru Avadanii <Alexandru.Avadanii@enea.com>
|
|
|
|
This reverts commit dff8f1a78db4a04dd30c2e0ac03a36cb0b198aa7.
Change-Id: I317e0f6aa7665c28035307149e6e7d42fdf508f3
|
|
|
|
This reverts commit 376c76d05862ed38edeb217cd9480807571491a7.
Change-Id: Ice749599287221af71128ec466894fdb8106d2ec
|
|
Slight rebase after upstream change [1].
FIXME: While looking at [1], I noticed that Armband is not
using the proper fuel-mirror version due to our divergent
patch & build system. This should be fixed later.
[1] https://gerrit.opnfv.org/gerrit/#/c/20317/
Change-Id: I55798ba1004312cd737b6613744683437e98865a
Signed-off-by: Alexandru Avadanii <Alexandru.Avadanii@enea.com>
|
|
|
|
|
|
|
|
Upstream change: [1]
Slight rebase of our patches is now needed.
[1] https://gerrit.opnfv.org/gerrit/#/c/20043/
Change-Id: I955ae3c5c304d8c8da6e42680c81e9255487fb68
Signed-off-by: Alexandru Avadanii <Alexandru.Avadanii@enea.com>
|
|
Change-Id: Ibd522813a40a0b1c385c47e65f77b1a13c860269
Signed-off-by: Alexandru Avadanii <Alexandru.Avadanii@enea.com>
|
|
|
|
|
|
Change-Id: Ia54bd3955a3e3b534d8c4d8380de5870a106edc5
Signed-off-by: Alexandru Avadanii <Alexandru.Avadanii@enea.com>
|
|
Some eth drivers, such as those for APM X-Gene Mustang Board's NICs, do
not advertise operstate properly in sysfs, ie. it is advertised as
"unknown" whereas ethtool shows the NIC as fully functional with link
detected. This further affects "ip link show" output which is parsed
within _check_iface_ready() method.
Replace "ip link show" command with "ethtool" in order to obtain proper
results when operstate is unknown.
Launchpad bug: https://bugs.launchpad.net/fuel/+bug/1618480
Change-Id: If8e28d3720161d9ffe897b875af235561cf1062d
Signed-off-by: Stefan Sicleru <stefan.sicleru@enea.com>
|
|
|
|
Change-Id: Iaabd176f4032494c8ea40fd1264dafcac01b4a43
Signed-off-by: Alexandru Avadanii <Alexandru.Avadanii@enea.com>
|