Gerrit tests have intermittent failures so removing them from the build. Upstream gerrit
isn't running tests on their builds either and it's probably due to the same reason.
Change-Id: I2cb2afb104895a90aed821cb2c720b3061f27262
grenade needs a bigger timeout buffer between the outer and inner
timeouts, because it has to do setup_workspace twice. Saw a failure
in the gate because we took 9 minutes before starting grenade
script and later had a timeout.
Change-Id: Ibf4ff531b47b63fc3d982f82a1dad1997fc3680c
We are publishing at api.openstack.org top-level but the job
is copying to an api-ref subdirectory. Changing the Jenkins setup is the
better solution here than copying directly into publish-docs - it would
create a more complex script.
Remove macro upload-to-api-site and add the publisher directly to the
openstack-api-ref job since this is an exception from the other rules -
and this is the only job publishing to api.o.o.
Change-Id: Ib5511ed7cbf8dfc90cb42e295fe68f39aa1e0a8f
* modules/openstack_project/manifests/gerrit.pp: Use a match group
around the pattern for a Change-Id since Gerrit doesn't interpret $0
as a backreference to the full pattern.
Change-Id: I711e8a1d7bc0b2f3eff07425badf2cb2c1740bf6
ant dependency requires java 6 so alternatives was set to java 6 on an ant package
install. This change updates alternatives to point to java 7 after all packages
have been installed.
Change-Id: I7b8c975879dda085102a1d4740777d3d8c5f54ea
Add stackforge/cookbook-openstack-integration-test as a new cookbook. Once it is
available we will populate the necessary files. Also updated the Gerrit
notifications to include the ops-database, ops-messaging and orchestration
cookbooks which were missing.
Change-Id: If6c7a25becc7642cd5ec07dce189947106698fc8
Addresses: blueprint integration-testing-for-havana
* .../templates/nodepool/nodepool.yaml.erb: Increase max-servers in
rax-ord to 92, since the quota there is 100.
Change-Id: Ie3fb939959274bdd723047bdd4fe28401d913561
* .../templates/nodepool/nodepool.yaml.erb: Since we still have a
handful of long-lived static slaves for sensitive jobs (proposal,
salt-trigger, mirrorXX) along with a couple dev slaves (precise,
centos6) and the remaining four precise-py3k slaves, the effective
capacity for nodepool in rax-dfw is 181. Cut this by a few more so
we have room for template servers during image updates.
Change-Id: I7cdfdba0665fff822fff77c31fa20a6585da411d
* .../nodepool/scripts/cache_git_repos.py: Since the projects list
is not being retrieved as part of a git repository, use HTTPS
protocol rather than HTTP so it will be consistent with the way we
retrieve the bootstrap script in prepare_node.sh.
* .../nodepool/scripts/prepare_node.sh: Switch the config clone
operation from the Gerrit server to the Git server farm, since the
former does not support git protocol. Also make it a shallow clone
since we just need the files.
Change-Id: I1dc0697bb3f9832d180b6b1ebffaa563bd886281
The ant build tool has a package and runtime dependency on java 6. The runtime dependency
can be changed by setting JAVA_HOME. This change sets the ant runtime dependency to
java 7.
Change-Id: I5d2ac7fb3dc7bdb390edcc728b86a84c946f8752
To try to understand more about what would cause a zuul-geard
disconnect. Only keep 3 days of logs because they are BIG.
Change-Id: Iba9a8a9c21d82ab4e3323db8c561f8de8c18864f
Previously, nodepool's prep scripts would not abort when
encountering git cloning errors. Now we retry once on failure and
then abort if it fails a second time. Also the script now makes sure
the clone actually results in a functional copy of each repository
by attempting to list branches and reset to HEAD.
Change-Id: I84565e4e90b34c6a9b8c86754efdde137cda08b2
Since devstack-gate no longer relies on the workspace-cache
directory, this is an unnecessary waste of effort.
Change-Id: Ic9ef8b269d0184f5891d355901bbfb29f7273670
For performance reasons--using git protocol instead of HTTP should
speed up cloning significantly on CentOS 6 in particular.
Change-Id: I37fadb1366eab5d66ec8e95945bea9c2c0251fcb
this is the isolated & parallel neutron job, and it now succeeds
(at least some of the time). So put it into check in a non-voting
way to gather enough data to make it voting.
remove the gate- version on the experimental queues, as we'd be
running on check all the time anyway.
Change-Id: I151d6960de32be1974052bf6f839901305dff403
If a change had a -1 verified vote on it from jenkins zuul would not
enqueue that change for check or gate testing when that change was
approved. Zuul should enqueue the change for check testing then move it
to the gate pipeline if a verified +1 comes back from check.
Change-Id: Icdcc4878a8289a2088c85afdfc72c56d4fd3a55a
There was an error in the conditional due to spacing (no space after the quotes). To fix
it I decided to replace the conditional with simpler 'mkdir -p' command.
Change-Id: I0f83e2823c69c5f9e3427e0925335e54e9b6c2a0
These are projects we'll want symmetric gating on eventually, and a
new pipeline allows us to work up tests and make sure things are
working properly in advance of having a sufficiently high uptime cloud
with fast enough tests to be in the main experimental queue.
Change-Id: Ic6369910ba162648608e17ec2c4057dc982120f5
