Nodepool expects the image updating cron to be called image-update. The
update-image cron that we had ends up being a noop and the default
image-update time is used instead. We want to change the time the image
updating happens which requires we use the correct name.
Change-Id: Ibc84d28705a0bebf95cc105e658941def9108bf7
Since we use trunk clients in the stable jobs, we should run stable jobs
on client changes. Add a client-stable-jobs template to run stable/havana
and stable/icehouse jobs on client changes.
Change-Id: I126d81aac5f5c165fcd6217ba7331967e8a3d345
We are seeing a race failure in check-tempest-dsvm-postgres-full jobs
where the error message shows up in the screen-n-api-meta log but we can
only get at it from screen-n-api logs (or console). We should index on
the screen-n-api-meta log so we can better filter the query.
Related-Bug: #921858
Change-Id: I1701ac83b2643d819245a7cdbbfb56cc4af12f5b
the -icehouse job branch limits were declared too late in the
layout.yaml causing them to have the unintended side effect of
making all -icehouse jobs voting (overriding earlier non-voting
stanzas).
move this to the top of the stack to ensure that we get the
behavior we were expecting.
Change-Id: I44ab42f2e405cc5c3e2cee5ba1f5afa85c39b8ee
tempest actually still has old branches, so we need to make sure
that we only run the branchless jobs on master and not on old
stable branches.
Closes-Bug: #1310368
Change-Id: I99e6c2185b40eb0ae32b56d18d0f7dad3e177f2a
The vcsrepo module doesn't really like the combo of latest and a
specific sha. It keeps grabbing it over and over again. Guess what?
The sha hasn't updated since the last time you looked for the same
sha.
Change-Id: I1a017278b83af11541794bf03644b3346e9869c2
The script was here for transition purposes, which are long-since
past. However, puppet runs it on every run. It's not hurting anything,
except for bunny rabbits.
Change-Id: I651cbf29b2bf137d470f4a29775622fea4e58835
The resource has _deb in it, which was left out.
Also, the java package name resource was replaced in only
one of two locations.
Change-Id: Ia47beeeff80f2cba1e7dfff0538156bd6464b9a1
java7-runtime-headless is an empty package that can never be
installed, so it attempts to install every time. Replace it with
openjdk-7-jre-headless, which is the real package.
Also, add a refreshonly on the exec for checking the diff, because
it runs every time as well.
Change-Id: I5195d27ae90cd7f9f3e20d92aca43d7b8d0e504a
We only want to start it when we've installed or upgraded.
The initial-init and init are designed to either one of them run,
one if there is no init script (initial install) the other if there
is (upgrade). The things they trigger should only run in response
to those actions.
Change-Id: I3bcaaed07bcf6239e053789b9b6241cbf652e7d4
Since ./exercise.sh is run via sudo -H, environment variables
are not available for exercises/rally.sh. Some variables are
needed to be preserved, so we run ./exercise.sh in post_test_hook
without sudo.
Change-Id: I4c147219037d9358e8abda32f2dca7bde5706f83
When we want to watch run_all happen, it's hard, because there is
no logging. To fix that - make there be some logging. Then, rotate
the logs.
Change-Id: I0eed7aeeec0ff21e58d57d6385cc59b74bbf31fb
When a thing is successful, we should not consider it to be a failure.
Doing so will cause it to have further mental anguish and feelings of
inadequacy. We don't support that sort of demeaning attitude in
OpenStack.
Change-Id: Ic92f5edf730297c200b916f21d0fe60dcab41d00
This reverts commit cdd09d30731678c99b5f20196145899af5cc1b35.
The gate for tripleo-image-elements is broken due to this commit. We
need to fix doc building and python26 before this will pass.
Change-Id: Icf02d05ff1e6f455b65c123a1091d4d32e6dbf1f