We must have forgotten to configure infra to upload releases of t-h-t to PyPi.
While t-h-t repo mainly consists of Heat templates, it also contains merge.py
script we are using.
Change-Id: I454476a0802928595b7ecb8577dfd67d19987b34
The last change* sadly missed a call to berks install and hence makes
strainer fail every time. This change introduces a berks install call
to a workspace-relative directory that is then passed to strainer.
To use strainer, we need chef-berkshelf-prep for each test suite, though.
*) I0353b8207abb116ea3849c04df39546f8c2625db
Change-Id: Ic10fc436fae2e5bcc05f9fae15e43b7601255471
The run-wheel.sh script echoed sha1 and md5 sums to the job console log
but did not write these hashes to files. Write these hashes to files
then cat them so that the hashes can be archived while still being
written to the console log.
Change-Id: If8d9eb8d730401938f8610bef355b5ca15723269
run-wheel.sh was deleting tarballs created by run-tarball.sh in the same
job. This meant Jenkins could not copy the tarballs and the tarball jobs
fail. Stop deleting tarballs and delete wheels instead as part of pre
wheel generation house keeping.
Change-Id: I4a850a289726ea6d25c72cf275fbba5aa60289b2
partial-bug: #1325456
The index page for status.openstack.org also has a link to
rechecks, menu has already been updated, update this link as
well and adjust text slightly.
Change-Id: I4c42894895e2ec6b6f4aca738c0d29ea93861c20
With the usage of openstack-doc-tools, markdown-docbook is not used
anymore, we can remove the script and the macro.
Change-Id: Id7375a50ee9183e0371c4089293b6515c9c0e240
After a recent change to api-site, the translation scripts tried to
translate the www directory which is wrong.
This patch unifies the translation setup of both propose and upstream
jobs into the common library and enhances the initial setup so that it
skips non-directories (like pom.xml) and special directories like www
and tools.
Change-Id: I9402d75ca9c8df67c62c56bbad582cb1166824c6
We had exclusions in the testing to allow for legacy issues and a
migration path. That's not cutting it any more and has allowed
a new -f line to slip in.
Be more strict.
Change-Id: I7c1debdd9307135efc68b402bf830a44c914d3fe
Env vars from separate script were exported only
for child process and were not available for other subprocesses.
https://jenkins07.openstack.org/job/gate-manila-tempest-dsvm-neutron/122/console
Set all external scripts to be called in parent process.
Because there is no need to use subprocesses.
Call of setup.sh is intended to be in pre_test_hook:
https://review.openstack.org/#/c/97003/
Also added 'export DEVSTACK_GATE_NEUTRON=1' to the config,
because name of job supposes using of neutron, so enable it
by default.
Change-Id: Ia211526d63c172b5c9f700d59c143ad78694348b
diskimage-builder does not run services that it installs, but this
is a problem when we lay down the resolv.conf file in puppet with
the service not running. So, put in a flag which defaults to true
(which is current behavior) that a dib invocation of the puppet
can turn off and then dib can plop the file down at the end of the
process.
Change-Id: I05d89ffacfdaf3563b8cb1460af12f114e1a0340
Just about every integrated project supports postgres (except swift),
but until now it wasn't part of the integrated-gate this means a
postgres breaking patch can be landed in the gate. This happened in heat
as documented by the bug below.
The integated-gate template is used on 35 repos while this patch removes
gate-tempest-dsvm-postgres in 20 places, so this patch adds a new gating
job to 15 repos.
Change-Id: Ieb1f162f9812917159314d328dde4f6c81f33e4c
Related-Bug: #1321093
what i meant was that we can just copy the same mechanism we're using to
run the minifier, instead of using the file object with ensure=>link
Fixed bug #1325117
Change-Id: I10cf4b4772e8d3fff869bfd638b2895b6e8980ea
The gate-config-puppet-apply test will try to run everything on a
bare-precise node, so even though this doesn't occur in production it is
being tested.
Change-Id: I13b03a7476e30cbb26dc820bb4cf5b56c4018669
We apparently didn't rebase all changes properly and some new scripts
ended up in jenkins/files/slave_scripts when they should all go into
openstack_project/files/slave_scripts. Move them now.
Change-Id: If6eec70f9da393193497d4d52d6a69a34cdb4db2
HP is turning off 1.0. Rip it out of our configs so that nodepool
doesn't go crazy when this happens.
Change-Id: I9aebbbbc7b78f2a057d4183568763a6d2d68ac25