ceph-docker-nightly: keep all logs forever

Signed-off-by: Andrew Schoen <aschoen@redhat.com>

Merge pull request #695 from ceph/libvirt-fixes

attempt to make libvirt slaves more stable

use libvirt util methods in all jobs that use libvirt

Signed-off-by: Andrew Schoen <aschoen@redhat.com>

scripts: add libvirt util functions to build_utils.sh

Signed-off-by: Andrew Schoen <aschoen@redhat.com>

ceph-installer-tests: restart libvirt services before running tests

Signed-off-by: Andrew Schoen <aschoen@redhat.com>

ceph-docker-prs: restart libvirt services before running tests

Signed-off-by: Andrew Schoen <aschoen@redhat.com>

ceph-docker-nightly: restart libvirt services before running tests

Signed-off-by: Andrew Schoen <aschoen@redhat.com>

ceph-ansible-prs: restart libvirt services before running tests

Signed-off-by: Andrew Schoen <aschoen@redhat.com>

ceph-installer-tests: remove old libvirt networks before running tests

Signed-off-by: Andrew Schoen <aschoen@redhat.com>

ceph-docker-prs: remove old libvirt networks before running tests

Signed-off-by: Andrew Schoen <aschoen@redhat.com>

ceph-docker-nightly: remove old libvirt networks before running tests

Signed-off-by: Andrew Schoen <aschoen@redhat.com>

Merge pull request #700 from ceph/wip-nfs-ganesha-version

nfs-ganesha: fix rpm version finder to work on dev and release candid…

nfs-ganesha: set VERSION correctly in rpm builds

$VERSION now set properly for "dev-X" & "rcX"
versions in rpm builds of nfs-ganesha.

Signed-off-by: Ali Maredia <amaredia@redhat.com>

Merge pull request #699 from ceph/ceph-ansible-scenario-fix

ceph-ansible-scenario: enable concurrent builds

ceph-ansible-scenario: enable concurrent builds

Signed-off-by: Andrew Schoen <aschoen@redhat.com>

Merge pull request #698 from ceph/ceph-ansible-scenario-fix

ceph-ansible-scenario: use include-raw instead of include-raw-escape

ceph-ansible-scenario: use include-raw instead of include-raw-escape

This job is not using a job-template and fails with the following error
when using include-raw-escape:

  syntax error near unexpected token `{{'

Signed-off-by: Andrew Schoen <aschoen@redhat.com>

Merge pull request #697 from ceph/ceph-ansible-scenario

ceph-ansible-scenario: create an individual scenario job

ceph-ansible-scenario: create an individual scenario job

Signed-off-by: Alfredo Deza <adeza@redhat.com>

Merge pull request #696 from ceph/disable-ceph-prs-arm

ceph-prs-arm: disable for now, needs to get worked on

ceph-prs-arm: disable for now, needs to get worked on

Signed-off-by: Alfredo Deza <adeza@redhat.com>

Merge pull request #693 from ceph/patched-jenkins

ansible: using forked version of python-jenkins until we can fix upstream

ansible: update examples with patched python-jenkins

Signed-off-by: Alfredo Deza <adeza@redhat.com>

ansible: using forked version of python-jenkins until we can fix upstream

Signed-off-by: Alfredo Deza <adeza@redhat.com>

Merge pull request #694 from ceph/rpm-install-mock

ceph-ansible/-installer: explicitly install mock

ceph-ansible/-installer: explicitly install mock

It appears that Yum is not installing "mock" when installing fedpkg
(with fedpkg-1.28-1.el7). Explicitly install it in our build steps.

Merge pull request #692 from ceph/slave-java-version

ansible: we now need java-1.8.0-openjdk on rpm nodes

ansible: we now need java-1.8.0-openjdk on rpm nodes

After the latest jenkins upgrade to 2.57 our centos slaves failed to
connect to jenkins, upgrading to a 1.8.0 version of java fixed the
issue.

Signed-off-by: Andrew Schoen <aschoen@redhat.com>

Merge pull request #691 from ceph/ganesha-json-fix

nfs-ganesha: add missing comma in JSON blob

nfs-ganesha: add missing comma in JSON blob

Signed-off-by: Alfredo Deza <adeza@redhat.com>

Merge pull request #689 from ceph/ceph-prs-arm

ceph-prs-arm: create a job to build arm on PRs

ceph-prs-arm: create a job to build arm on PRs

Signed-off-by: Alfredo Deza <adeza@redhat.com>

Merge pull request #690 from ceph/libvirt-hammer

ceph-ansible-prs: remove old libvirt networks before running tests

ceph-ansible-prs: remove old libvirt networks before running tests

Occasionally we deal with libvirt network issues on jenkins slaves that
have been used for a couple test runs, this attempts to avoid those.

Signed-off-by: Andrew Schoen <aschoen@redhat.com>

Merge pull request #688 from ceph/wip-bz-mill

re-organize the script

re-organize the script

Signed-off-by: Gregory Meno <gmeno@redhat.com>

Merge pull request #687 from ceph/wip-bz-mill

add config to run the template generator

add config to run the template generator

Signed-off-by: Gregory Meno <gmeno@redhat.com>

Merge pull request #686 from ceph/wip-bz-mill

see what is in the enviroment

use b-ranto's template trickto get the metrics map

Signed-off-by: Gregory Meno <gmeno@redhat.com>

see what is in the enviroment

Signed-off-by: Gregory Meno <gmeno@redhat.com>

Merge pull request #685 from ceph/wip-bz-mill

del the project -- don't need it

Merge branch 'master' into wip-bz-mill

del the project -- don't need it

Signed-off-by: Gregory Meno <gmeno@redhat.com>

Merge pull request #684 from ceph/luminous-trigger

ceph-dev-trigger: also build luminous

ceph-dev-trigger: also build luminous

Signed-off-by: Andrew Schoen <aschoen@redhat.com>

Merge pull request #683 from ceph/wip-bz-mill

adds a builder def.

adds a builder def.

Signed-off-by: Gregory Meno <gmeno@redhat.com>

Merge pull request #682 from ceph/wip-bz-mill

adds a job definition

adds a job definition

Signed-off-by: Gregory Meno <gmeno@redhat.com>

Merge pull request #681 from ceph/wip-bz-mill

adds a simple script that feeds bz counts into grafana

removing unnecessary teardown

Signed-off-by: Gregory Meno <gmeno@redhat.com>

adds a simple script that feeds bz counts into grafana

Signed-off-by: Gregory Meno <gmeno@redhat.com>

Merge pull request #679 from ceph/ceph-ansible-docs-prs

ceph-ansible-docs-prs: create a job for building docs on prs

ceph-ansible-docs-prs: create a job for building docs on prs

Signed-off-by: Alfredo Deza <adeza@redhat.com>

Merge pull request #678 from ceph/libvirt-networks

ceph-ansible-prs: destroy and undefine libvirt networks on teardown

ceph-ansible-prs: destroy and undefine libvirt networks on teardown

Signed-off-by: Andrew Schoen <aschoen@redhat.com>

Merge pull request #677 from ceph/libvirt-kill-networks

ceph-ansible-prs: ensure libvirt networks are destroyed after a job is completed

ceph-ansible-prs: ensure libvirt networks are destroyed after a job is completed

Signed-off-by: Alfredo Deza <adeza@redhat.com>

Merge pull request #676 from ceph/mkdir-typo

ceph-ansible-docs: fix a typo on mkdir

ceph-ansible-docs: fix a typo on mkdir

Signed-off-by: Alfredo Deza <adeza@redhat.com>

Merge pull request #675 from ceph/ceph-ansible-docs

ceph-ansible-docs: initial take on automatic docs for ceph-ansible

ceph-ansible-docs: initial take on automatic docs for ceph-ansible

Signed-off-by: Alfredo Deza <adeza@redhat.com>

Merge pull request #674 from ceph/pr-submodules-fetch

ceph-pr-submodules: explicitly fetch the target branch

ceph-pr-submodules: explicitly fetch the target branch

This ceph-pr-submodules job is still giving us false positives in some
cases.

We were comparing against a target branch (for example "origin/jewel"),
but we could not see the *exact* sha1 that Jenkins thought was
"origin/jewel". Sometimes this job would fail, and the failure hinted
that Jenkins might be comparing an older target branch.

Explicitly fetch the target branch immediately before doing the "git
log" operation, and print the target branch's sha1 ref to stdout in case
we need further debugging.

Merge pull request #673 from ceph/wip-nfs-ganesha

nfs-ganesha: remove trusty from axis

nfs-ganesha: remove trusty from axis

Signed-off-by: Ali Maredia <amaredia@redhat.com>

Merge pull request #672 from ceph/ceph-ansible-kraken

ceph-ansible-prs: run tests for jewel and kraken

ceph-ansible-prs: run tests for jewel and kraken

Signed-off-by: Andrew Schoen <aschoen@redhat.com>

Merge pull request #671 from ceph/wip-nfs-ganesha

nfs-ganesha: change shaman polling time to 20 mins

nfs-ganesha: change shaman polling time to 20 mins

Certain Trusty builds were failing because the
repos weren't available until 11-15 mins after
the job was run

Signed-off-by: Ali Maredia <amaredia@redhat.com>

Merge pull request #670 from ceph/wip-nfs-ganesha

nfs-ganesha: fix build_deb shaman polling loop

nfs-ganesha: fix build_deb shaman polling loop exit status

The curl command sometimes returns 504, since the
chacra repository isn't ready yet. That exit
status stops the rest of the script from running
instead of running through the loop for $TIME_LIMIT
and rerunning curl every $INTERVAL.

This fix
ensures the exit status is always true.

Signed-off-by: Ali Maredia <amaredia@redhat.com>

Merge pull request #669 from ceph/wip-nfs-ganesha

nfs-ganesha: add missing quotation mark

nfs-ganesha: add missing quotation mark

Signed-off-by: Ali Maredia <amaredia@redhat.com>

Merge pull request #667 from ceph/pbuilder-override-config

reconfigure apt during every "pbuilder update"

reconfigure apt during every "pbuilder update"

The nfs-ganesha pbuilder setup code configures the pbuilder caches to
point at repos from shaman. Since all Jenkins jobs uses the same
pbuilder cache tarballs, this means that these shaman repositories will
still be enabled in the next jobs that run on a slave.

1) nfs-ganesha job runs on a slave
2) nfs-ganesha job configures pbuilder cache to use shaman repo
3) nfs-ganesha job finishes
4) ceph-build job runs on same slave slave
5) ceph-build job uses same pbuilder cache with shaman repo configured

In addition to polluting the pbuilder environment, this can also lead to
pbuilder failures when shaman/chacra garbage-collects old repositories,
because Apt wants to download repo metadata from old ceph repos that
are no longer available.

Force every job that runs "pbuilder update" to also run with
"--override-config" in order to wipe out any apt repositories that may
have been configured in the cache.

Merge pull request #666 from ceph/wip-nfs-ganesha

nfs-ganesha: put flavor after chacra_endpoint

nfs-ganesha: add flavor to chacra endpoints

Signed-off-by: Ali Maredia <amaredia@redhat.com>

Merge pull request #665 from dmick/wip-submodule-check

ceph-pr-submodules: fix submodule check.

ceph-pr-submodules: fix submodule check.

Prior code never generated any output at all.

Fixes: http://tracker.ceph.com/issues/19244
Signed-off-by: Dan Mick <dan.mick@redhat.com>

Merge pull request #664 from dmick/wip-avoid-null-doc-build

ceph-docs: Verify push contains docs files; don't build if not

ceph-docs: Verify push contains docs files; don't build if not

Signed-off-by: Dan Mick <dan.mick@redhat.com>

Merge pull request #663 from ceph/build-utils-ssl-verify

build_utils: enable SSL verification

build_utils: enable SSL verification

All our chacra endpoints (https://*.chacra.ceph.com) use HTTPS certs
signed by LetsEncrypt now. There's no need to disable SSL verification
here any more.

(This conceptually reverts the change in
f5e28bdeb675f6e245c26ae7d0e7f662ac1ed0f1).

Signed-off-by: Ken Dreyer <kdreyer@redhat.com>

Merge pull request #662 from ktdreyer/nfs-ganesha-rm-credential

nfs-ganesha: rm SSH credential

nfs-ganesha: rm SSH credential

This job doesn't use SSH, so there's no need to make this credential
available to the job.

Merge pull request #661 from ceph/wip-nfs-ganesha

nfs-ganesha: add flavors and assing BRANCH to NFS_GANESHA_BRANCH

nfs-ganesha: add flavors and assing BRANCH to NFS_GANESHA_BRANCH

Signed-off-by: Ali Maredia <amaredia@redhat.com>

Merge pull request #660 from dmick/master

setup_pbuilder: fix boolean test

setup_pbuilder: fix boolean test

three times

Signed-off-by: Dan Mick <dan.mick@redhat.com>

Merge pull request #659 from ceph/wip-nfs-ganesha-ceph-dev-build-trigger

nfs-ganesha: add property files from ceph-dev-build to nfs-ganesha job

nfs-ganesha: add property files from ceph-dev-build to nfs-ganesha job

Signed-off-by: Ali Maredia <amaredia@redhat.com>

Merge pull request #658 from ceph/jenkins-nginx-timeout

ansible: double the time for nginx-to-jenkins for timeouts

ansible: double the time for nginx-to-jenkins for timeouts

Signed-off-by: Alfredo Deza <adeza@redhat.com>

Merge pull request #657 from ceph/wip-nfs-ganesha-ceph-build-trigger

nfs-ganesha: $CEPH_SHA1 now takes $SHA1 from ceph-dev-build job

nfs-ganesha: $CEPH_SHA1 now takes $SHA1 from ceph-dev-build job

Signed-off-by: Ali Maredia <amaredia@redhat.com>

Merge pull request #655 from ceph/ceph-ansible-prs

ceph-ansible-prs: adds the ansible2.2-update_cluster scenario

ceph-ansible-prs: adds the ansible2.2-update_cluster scenario

Signed-off-by: Andrew Schoen <aschoen@redhat.com>

Merge pull request #654 from ceph/wip-nfs-ganesha-ceph-dev-trigger

nfs-ganesha: trigger builds from ceph-dev-build