]> git.apps.os.sepia.ceph.com Git - ceph-ansible.git/log
ceph-ansible.git
8 years agoFix the mons running check to use group name var 1168/head
Logan V [Thu, 8 Dec 2016 19:16:02 +0000 (13:16 -0600)]
Fix the mons running check to use group name var

mon_group_name variable can be used to override mons group, but
this task assumes the group is always 'mons'. So we need to use
the var to find the group name instead.

8 years agoMerge pull request #1157 from ceph/ceph-fsid
Sébastien Han [Thu, 8 Dec 2016 17:55:21 +0000 (18:55 +0100)]
Merge pull request #1157 from ceph/ceph-fsid

common: do not regenerate a cluster fsid if cluster exits

8 years agoMerge pull request #1162 from ceph/timeout-collect-keys
Andrew Schoen [Thu, 8 Dec 2016 16:28:28 +0000 (10:28 -0600)]
Merge pull request #1162 from ceph/timeout-collect-keys

mon: remove collect keys

8 years agoMerge pull request #1150 from ceph/testinfra
Alfredo Deza [Thu, 8 Dec 2016 15:22:24 +0000 (10:22 -0500)]
Merge pull request #1150 from ceph/testinfra

tests: use testinfra to verify cluster setup in our CI tests

8 years agoMerge pull request #1164 from guits/master
Sébastien Han [Thu, 8 Dec 2016 10:24:11 +0000 (11:24 +0100)]
Merge pull request #1164 from guits/master

git: update gitignore

8 years agogit: update gitignore 1164/head
Guillaume Abrioux [Thu, 8 Dec 2016 07:17:35 +0000 (08:17 +0100)]
git: update gitignore

* ignore `*.retry` files in general

Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
8 years agotests: adds an osds parameter to the node fixture 1150/head
Andrew Schoen [Thu, 8 Dec 2016 01:24:40 +0000 (19:24 -0600)]
tests: adds an osds parameter to the node fixture

The osds are named differently for systemd in containerized deployments
so this new parameter is used to make that change transparent in the
tests.

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
8 years agotests: mark tests with no_docker that fail on containerized deployments
Andrew Schoen [Thu, 8 Dec 2016 01:19:43 +0000 (19:19 -0600)]
tests: mark tests with no_docker that fail on containerized deployments

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
8 years agotests: adds a docker property to the node fixture
Andrew Schoen [Thu, 8 Dec 2016 01:17:09 +0000 (19:17 -0600)]
tests: adds a docker property to the node fixture

This is used to skip tests that will not work with containerized
deployment or atomic hosts.

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
8 years agotests: adds a docker: true var to the docker_cluster scenario
Andrew Schoen [Thu, 8 Dec 2016 01:15:29 +0000 (19:15 -0600)]
tests: adds a docker: true var to the docker_cluster scenario

This lets the CI tests know that this scenario is using docker and
should skip certain tests.

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
8 years agodocker: fix ceph.conf generation for multiple mons
Andrew Schoen [Wed, 7 Dec 2016 23:02:02 +0000 (17:02 -0600)]
docker: fix ceph.conf generation for multiple mons

Before this patch only the address for the first mon would show
in the ceph.conf even if there were multiple mons in the inventory.

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
8 years agotests: account for multiple osd hosts when testing if all are up and in
Andrew Schoen [Wed, 7 Dec 2016 20:31:31 +0000 (14:31 -0600)]
tests: account for multiple osd hosts when testing if all are up and in

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
8 years agotests: ensure all osds are listening on the cluster network
Andrew Schoen [Tue, 6 Dec 2016 22:52:02 +0000 (16:52 -0600)]
tests: ensure all osds are listening on the cluster network

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
8 years agotests: test that all osds are listening on the public network
Andrew Schoen [Tue, 6 Dec 2016 22:46:42 +0000 (16:46 -0600)]
tests: test that all osds are listening on the public network

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
8 years agotests: add num_devices to node fixture
Andrew Schoen [Wed, 7 Dec 2016 16:32:00 +0000 (10:32 -0600)]
tests: add num_devices to node fixture

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
8 years agomon: remove collect keys 1162/head
Sébastien Han [Wed, 7 Dec 2016 13:36:24 +0000 (14:36 +0100)]
mon: remove collect keys

Once the monitor process starts it will also trigger `ceph-create-keys`
which will collect the admin key and bootstrap keys. We used to force
this command because we were having issues on some distros like centos
7.0 and 7.1 not triggering this. This is fixed on centos 7.2 and not an
issue on ubuntu 14.04 or 16.04 so we can remove this task. If the
monitor hangs or fails to start the playbook will fail right after at
the "wait for client.admin key exists" task after 300sec.

Closes: #1161
Signed-off-by: Sébastien Han <seb@redhat.com>
8 years agocommon: do not regenerate a cluster fsid if cluster exists 1157/head
Sébastien Han [Tue, 6 Dec 2016 14:59:49 +0000 (15:59 +0100)]
common: do not regenerate a cluster fsid if cluster exists

This commit solves the situation where you lost your fetch directory and
you are running ansible against an existing cluster. Since no fetch
directory is present the file containing the fsid doesn't exist so we
are creating a new one. Later the ceph.conf gets updated with a wrong
fsid which causes problems for clients and ceph processes.

Closes: #1148
Signed-off-by: Sébastien Han <seb@redhat.com>
8 years agotests: adds a comment to the setup.yml playbook for clarification
Andrew Schoen [Tue, 6 Dec 2016 20:24:36 +0000 (14:24 -0600)]
tests: adds a comment to the setup.yml playbook for clarification

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
8 years agotests: skip installing net-tools on atomic hosts
Andrew Schoen [Tue, 6 Dec 2016 20:23:12 +0000 (14:23 -0600)]
tests: skip installing net-tools on atomic hosts

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
8 years agotests: dmcrypt_journal_collocation should run journal collocation tests
Andrew Schoen [Tue, 6 Dec 2016 20:13:50 +0000 (14:13 -0600)]
tests: dmcrypt_journal_collocation should run journal collocation tests

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
8 years agotests: use cluster_name in tests when needed
Andrew Schoen [Tue, 6 Dec 2016 20:07:40 +0000 (14:07 -0600)]
tests: use cluster_name in tests when needed

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
8 years agotests: use node["conf_path"] instead of hardcoding the path in tests
Andrew Schoen [Tue, 6 Dec 2016 19:39:08 +0000 (13:39 -0600)]
tests: use node["conf_path"] instead of hardcoding the path in tests

This also accounts for the cluster having a custom cluster name

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
8 years agotests: add cluster_name and conf_path to the node fixture
Andrew Schoen [Tue, 6 Dec 2016 19:32:28 +0000 (13:32 -0600)]
tests: add cluster_name and conf_path to the node fixture

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
8 years agotests: include the number of mons as a attribute on node
Andrew Schoen [Tue, 6 Dec 2016 16:55:20 +0000 (10:55 -0600)]
tests: include the number of mons as a attribute on node

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
8 years agotests: update skip messages in conftest.py
Andrew Schoen [Tue, 6 Dec 2016 16:50:54 +0000 (10:50 -0600)]
tests: update skip messages in conftest.py

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
8 years agotests: rename vars to ansible_vars in conftest.node
Andrew Schoen [Tue, 6 Dec 2016 15:58:01 +0000 (09:58 -0600)]
tests: rename vars to ansible_vars in conftest.node

This is to avoid redefining the builtin vars()

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
8 years agotests: rename CephNode to node and add a bit of documentation
Andrew Schoen [Tue, 6 Dec 2016 15:54:50 +0000 (09:54 -0600)]
tests: rename CephNode to node and add a bit of documentation

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
8 years agotests: with testinfra we don't need scenario.py files anymore
Andrew Schoen [Mon, 5 Dec 2016 22:27:19 +0000 (16:27 -0600)]
tests: with testinfra we don't need scenario.py files anymore

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
8 years agotests: tests for collocated journals
Andrew Schoen [Mon, 5 Dec 2016 22:23:15 +0000 (16:23 -0600)]
tests: tests for collocated journals

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
8 years agotests: add a journal_collocation tag to collocated journal tests
Andrew Schoen [Mon, 5 Dec 2016 22:21:52 +0000 (16:21 -0600)]
tests: add a journal_collocation tag to collocated journal tests

The test will be skipped if journal_collocation is not True in the
group_vars for the scenario.

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
8 years agotests: ensure ceph-mon is installed on mons
Andrew Schoen [Mon, 5 Dec 2016 22:21:10 +0000 (16:21 -0600)]
tests: ensure ceph-mon is installed on mons

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
8 years agotests: make sure all osds are mounted
Andrew Schoen [Mon, 5 Dec 2016 21:54:11 +0000 (15:54 -0600)]
tests: make sure all osds are mounted

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
8 years agotests: adds some basic tests for OSD nodes
Andrew Schoen [Mon, 5 Dec 2016 21:05:23 +0000 (15:05 -0600)]
tests: adds some basic tests for OSD nodes

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
8 years agotesting: CephNode should return osd_ids
Andrew Schoen [Mon, 5 Dec 2016 21:03:44 +0000 (15:03 -0600)]
testing: CephNode should return osd_ids

This is an empty list on anything but an OSD node. These ids will be
used for further OSD testing.

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
8 years agotests: remove test files that are not used anymore
Andrew Schoen [Mon, 5 Dec 2016 18:07:53 +0000 (12:07 -0600)]
tests: remove test files that are not used anymore

These tests have been replaced by the ones in test_install.py

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
8 years agotests: test ceph.conf using testinfra modules
Andrew Schoen [Mon, 5 Dec 2016 18:06:46 +0000 (12:06 -0600)]
tests: test ceph.conf using testinfra modules

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
8 years agotests: ensure tests marked with 'all' run on every node
Andrew Schoen [Mon, 5 Dec 2016 16:07:04 +0000 (10:07 -0600)]
tests: ensure tests marked with 'all' run on every node

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
8 years agotests: the CephNode fixture should provide the cluster subnet
Andrew Schoen [Mon, 5 Dec 2016 15:47:33 +0000 (09:47 -0600)]
tests: the CephNode fixture should provide the cluster subnet

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
8 years agotests: tests mon services to ensure they are running correctly
Andrew Schoen [Sun, 4 Dec 2016 02:10:35 +0000 (20:10 -0600)]
tests: tests mon services to ensure they are running correctly

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
8 years agotests: run testinfra with --sudo
Andrew Schoen [Sun, 4 Dec 2016 02:09:37 +0000 (20:09 -0600)]
tests: run testinfra with --sudo

It needs this so that the modules can use sudo

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
8 years agotests: run the setup.yml playbook with tox
Andrew Schoen [Sun, 4 Dec 2016 02:08:53 +0000 (20:08 -0600)]
tests: run the setup.yml playbook with tox

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
8 years agotests: include a playbook to install net-tools for socket checks
Andrew Schoen [Sun, 4 Dec 2016 02:02:05 +0000 (20:02 -0600)]
tests: include a playbook to install net-tools for socket checks

This playbook could be used in the future to install anything else we
need on these nodes for testing purposes.

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
8 years agotests: return address and ansible vars from CephNode
Andrew Schoen [Sun, 4 Dec 2016 02:01:30 +0000 (20:01 -0600)]
tests: return address and ansible vars from CephNode

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
8 years agotests: port tests/functional/test_install.py to use testinfra fixtures
Andrew Schoen [Sat, 3 Dec 2016 14:10:54 +0000 (08:10 -0600)]
tests: port tests/functional/test_install.py to use testinfra fixtures

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
8 years agotests: change conftest.py to support dynamic markers and test selection
Andrew Schoen [Sat, 3 Dec 2016 14:07:09 +0000 (08:07 -0600)]
tests: change conftest.py to support dynamic markers and test selection

This applies correct markers to tests depending on their file path.
Tests that live at the root of tests/functional/tests will be given a
marker of 'all' and apply to any type of ceph node.

Tests that live in tests/functional/tests/mon will get the
'mons' marker and will only run for nodes in the 'mons' group.

Tests that live in tests/functional/tests/osd will get the 'osds'
marker and will only run for nodes in the 'osds' group.

All tests must use the CephNode fixture for this to work and be
parameterized by testinfra.

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
8 years agotests: use testinfra in tox.ini instead of pytest directly
Andrew Schoen [Sat, 3 Dec 2016 14:06:20 +0000 (08:06 -0600)]
tests: use testinfra in tox.ini instead of pytest directly

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
8 years agoMerge pull request #1155 from ceph/take-over-fix
Sébastien Han [Tue, 6 Dec 2016 15:04:52 +0000 (16:04 +0100)]
Merge pull request #1155 from ceph/take-over-fix

take-over: use more precise ceph.conf detection

8 years agotake-over: use more precise ceph.conf detection 1155/head
Sébastien Han [Tue, 6 Dec 2016 14:02:49 +0000 (15:02 +0100)]
take-over: use more precise ceph.conf detection

Prior to this patch we were just looking for any *.conf file which
sometimes could results in multiple matches. The new command looks for a
.conf file that must contain [global] and 'fsid' patterns. This will
definitely get us the ceph.conf file. We can not directly use ceph.conf
because of a different cluster name.

Signed-off-by: Sébastien Han <seb@redhat.com>
8 years agoMerge pull request #1152 from ceph/gitignore-travis
Sébastien Han [Tue, 6 Dec 2016 09:19:23 +0000 (10:19 +0100)]
Merge pull request #1152 from ceph/gitignore-travis

Gitignore travis

8 years agogit: update gitignore 1152/head
Sébastien Han [Thu, 1 Dec 2016 15:34:03 +0000 (16:34 +0100)]
git: update gitignore

* ignore yml files in general
* refactor based on commit f8e043b6ea5ac4e886532d4f2f675c507b44b955 that
changed directory layouts

Signed-off-by: Sébastien Han <seb@redhat.com>
(cherry picked from commit ec5c6f5da566611c4e0b88f925cbd26dc90368d6)

8 years agoci: remove travis
Sébastien Han [Thu, 1 Dec 2016 15:28:56 +0000 (16:28 +0100)]
ci: remove travis

since now we have the ceph CI taking care of the syntax check we don't
need travis anymore. Thanks Travis you served us well!

Signed-off-by: Sébastien Han <seb@redhat.com>
(cherry picked from commit 9fbe65cb72e01c586af3a27788c8e3e9549e0abf)

8 years agoMerge pull request #1144 from yanshanchen/fix-shrink-osd-regex
Sébastien Han [Mon, 5 Dec 2016 09:14:58 +0000 (10:14 +0100)]
Merge pull request #1144 from yanshanchen/fix-shrink-osd-regex

the pattern in regex_replace is not right when the pid of osd is bigger  than 9999

8 years agothis patch fix the regex pattern in infrastructure-playbooks/shrink-osd.yml when... 1144/head
chenyanshan [Mon, 5 Dec 2016 05:40:38 +0000 (13:40 +0800)]
this patch fix the regex pattern in infrastructure-playbooks/shrink-osd.yml  when the  osd's pid num is bigger than 9999

Signed-off-by: chenyanshan <yanshanchen@139.com>
8 years agoMerge pull request #1141 from ceph/more-mons
Alfredo Deza [Fri, 2 Dec 2016 15:27:05 +0000 (10:27 -0500)]
Merge pull request #1141 from ceph/more-mons

centos7_cluster and xenial_cluster scenarios should use 3 mons

8 years agotests: make centos7_cluster use three mons 1141/head
Andrew Schoen [Thu, 1 Dec 2016 23:27:09 +0000 (17:27 -0600)]
tests: make centos7_cluster use three mons

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
8 years agotests: rename the directory for centos7_cluster
Andrew Schoen [Thu, 1 Dec 2016 22:59:27 +0000 (16:59 -0600)]
tests: rename the directory for centos7_cluster

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
8 years agotests: removes unused xenial_conf_tests and xenial_mon_osd scenarios
Andrew Schoen [Thu, 1 Dec 2016 22:56:06 +0000 (16:56 -0600)]
tests: removes unused xenial_conf_tests and xenial_mon_osd scenarios

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
8 years agotests: rename the directory for xenial_cluster
Andrew Schoen [Thu, 1 Dec 2016 22:48:06 +0000 (16:48 -0600)]
tests: rename the directory for xenial_cluster

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
8 years agotests: add 2 more mons to xenial_cluster
Andrew Schoen [Thu, 1 Dec 2016 22:46:05 +0000 (16:46 -0600)]
tests: add 2 more mons to xenial_cluster

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
8 years agoMerge pull request #1138 from ceph/docker-scenario
Ivan Font [Thu, 1 Dec 2016 21:56:54 +0000 (13:56 -0800)]
Merge pull request #1138 from ceph/docker-scenario

tests: adds a docker_cluster scenario

8 years agotests: remove unused group_vars/osds from docker_cluster 1138/head
Andrew Schoen [Thu, 1 Dec 2016 20:50:21 +0000 (14:50 -0600)]
tests: remove unused group_vars/osds from docker_cluster

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
8 years agotests: docker_cluster does not need os_tuning_params
Andrew Schoen [Thu, 1 Dec 2016 20:49:31 +0000 (14:49 -0600)]
tests: docker_cluster does not need os_tuning_params

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
8 years agotests: docker_cluster should run conf_tests
Andrew Schoen [Thu, 1 Dec 2016 17:12:46 +0000 (11:12 -0600)]
tests: docker_cluster should run conf_tests

This also makes the conf_tests take the subnet as input
so multiple scenarios on differing subnets can use these tests.

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
8 years agotests: use journal_size in ceph_osd_docker_extra_env for docker_cluster
Andrew Schoen [Thu, 1 Dec 2016 16:27:48 +0000 (10:27 -0600)]
tests: use journal_size in ceph_osd_docker_extra_env for docker_cluster

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
8 years agotests: remove unsued options from docker_cluster vagrant variables
Andrew Schoen [Thu, 1 Dec 2016 16:24:30 +0000 (10:24 -0600)]
tests: remove unsued options from docker_cluster vagrant variables

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
8 years agotests: docker_cluster should use 3 mons
Andrew Schoen [Thu, 1 Dec 2016 16:21:29 +0000 (10:21 -0600)]
tests: docker_cluster should use 3 mons

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
8 years agotests: docker_cluster should use the site-docker.yml playbook
Andrew Schoen [Thu, 1 Dec 2016 16:19:19 +0000 (10:19 -0600)]
tests: docker_cluster should use the site-docker.yml playbook

This commits allows for scenarios to pick their own playbook while
defaulting to use site.yml.sample.

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
8 years agoceph-rgw: default rgw_config_keys in defaults/main.yml
Andrew Schoen [Wed, 30 Nov 2016 22:47:51 +0000 (16:47 -0600)]
ceph-rgw: default rgw_config_keys in defaults/main.yml

This avoid an issue with it being undefined if nfs_obj_gw is
set to False.

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
8 years agotests: update the group_vars for docker_cluster
Andrew Schoen [Wed, 30 Nov 2016 21:21:12 +0000 (15:21 -0600)]
tests: update the group_vars for docker_cluster

Fixes the monitor_interface and sets osd scenario config
options so the mandatory checks pass.

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
8 years agoceph-rgw: no need to use playbook_dir when fetching configs for docker
Andrew Schoen [Wed, 30 Nov 2016 21:18:36 +0000 (15:18 -0600)]
ceph-rgw: no need to use playbook_dir when fetching configs for docker

This avoids a bug when fetch_directory is not a relative to the playbook
directory.

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
8 years agoceph-osd: no need to use playbook_dir when fetching configs for docker
Andrew Schoen [Wed, 30 Nov 2016 21:11:24 +0000 (15:11 -0600)]
ceph-osd: no need to use playbook_dir when fetching configs for docker

This causes a bug when fetch_directory is not relative to the playbook
directory.

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
8 years agoceph-mon: no need to include playbook_dir when fetching configs
Andrew Schoen [Wed, 30 Nov 2016 21:08:42 +0000 (15:08 -0600)]
ceph-mon: no need to include playbook_dir when fetching configs

This causes a bug when fetch_directory is not a relative path to the
playbook location.

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
8 years agoceph-mon: adds static: no to an include to avoid an ansible bug
Andrew Schoen [Wed, 30 Nov 2016 21:06:06 +0000 (15:06 -0600)]
ceph-mon: adds static: no to an include to avoid an ansible bug

Adding that avoids this bug:

https://github.com/ansible/ansible/issues/18206

Without that you'll get failures like:

TASK [ceph-mon : set keys permissions]
*****************************************
task path:
/home/andrewschoen/ceph-ansible/roles/ceph-mon/tasks/ceph_keys.yml:31
fatal: [mon0]: FAILED! => {"failed": true, "msg": "'dict object' has no attribute 'stdout_lines'"}

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
8 years agotests: adds a docker_cluster scenario
Andrew Schoen [Wed, 30 Nov 2016 18:20:07 +0000 (12:20 -0600)]
tests: adds a docker_cluster scenario

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
8 years agotests: change the subnet on dmcrypt-journal-collocation
Andrew Schoen [Mon, 28 Nov 2016 21:43:51 +0000 (15:43 -0600)]
tests: change the subnet on dmcrypt-journal-collocation

It was currently set conflicting with dmcrypt-journal

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
8 years agoMerge pull request #1139 from ceph/last-cleanup-apache
Sébastien Han [Thu, 1 Dec 2016 10:04:48 +0000 (11:04 +0100)]
Merge pull request #1139 from ceph/last-cleanup-apache

common: remove apache file

8 years agocommon: remove apache file 1139/head
Sébastien Han [Thu, 1 Dec 2016 10:02:31 +0000 (11:02 +0100)]
common: remove apache file

we recently dropped the support for apache with rgw, so this commit
removes the last remaining file.

Signed-off-by: Sébastien Han <seb@redhat.com>
8 years agoMerge pull request #1135 from guits/master
Sébastien Han [Wed, 30 Nov 2016 16:17:48 +0000 (17:17 +0100)]
Merge pull request #1135 from guits/master

Add gluster nfs ganesha repo only for Ubuntu

8 years agoAdd gluster nfs ganesha repo only for Ubuntu 1135/head
Guillaume Abrioux [Wed, 30 Nov 2016 15:30:43 +0000 (16:30 +0100)]
Add gluster nfs ganesha repo only for Ubuntu

According to the official documentation, adding PPA is not available for Debian.

8 years agoMerge pull request #1125 from guits/master
Sébastien Han [Wed, 30 Nov 2016 14:50:23 +0000 (15:50 +0100)]
Merge pull request #1125 from guits/master

Use 'package' module instead of yum, apt and dnf

8 years agoMerge pull request #1131 from guits/refact_code
Sébastien Han [Wed, 30 Nov 2016 14:46:15 +0000 (15:46 +0100)]
Merge pull request #1131 from guits/refact_code

Refact temporary vars in ceph-common defaults.

8 years agoRefact temporary vars in ceph-common defaults. 1131/head
Guillaume Abrioux [Tue, 29 Nov 2016 17:14:03 +0000 (18:14 +0100)]
Refact temporary vars in ceph-common defaults.

These variables were defined here to be sure that
`roles/ceph-common/tasks/checks/check_mandatory_vars.yml` has all variables defined.

8 years agoMerge pull request #1132 from ktdreyer/no-ceph-extra
Andrew Schoen [Tue, 29 Nov 2016 22:21:51 +0000 (16:21 -0600)]
Merge pull request #1132 from ktdreyer/no-ceph-extra

rm ceph-extra.repo and redhat_distro_ceph_extra variable

8 years agorm ceph-extra.repo and redhat_distro_ceph_extra variable 1132/head
Ken Dreyer [Tue, 29 Nov 2016 20:02:58 +0000 (13:02 -0700)]
rm ceph-extra.repo and redhat_distro_ceph_extra variable

We removed the "apache" setting for "radosgw_frontend" in
adfdf6871e1172b542342064c13147789f39571a.

As part of that change, we removed the final references to
ceph-extra.repo, but I failed to clean up this file itself.

Now that nothing uses this file, delete it.

This file contained the sole reference to redhat_distro_ceph_extra, so
we can drop that variable as well.

8 years agoMerge pull request #1130 from ceph/colocated-tests
Andrew Schoen [Tue, 29 Nov 2016 21:40:25 +0000 (15:40 -0600)]
Merge pull request #1130 from ceph/colocated-tests

initial pass on tests for collocated journals

8 years agotests: use underscores to prevent tox combinatorial issues 1130/head
Alfredo Deza [Tue, 29 Nov 2016 20:01:31 +0000 (15:01 -0500)]
tests: use underscores to prevent tox combinatorial issues

Signed-off-by: Alfredo Deza <adeza@redhat.com>
Resolves: colocated#tests

8 years agoUse 'package' module instead of yum, apt and dnf 1125/head
Guillaume Abrioux [Mon, 28 Nov 2016 14:16:56 +0000 (15:16 +0100)]
Use 'package' module instead of yum, apt and dnf

Refactor the code using 'package' module

Fix Issue #520

(However it doesn't cover all cases because some cases are not refactorable.
Ex: because of diverging packages name between distribution)

8 years agotests: specify the devices for OSDs in the scenario for collocation
Alfredo Deza [Tue, 29 Nov 2016 15:05:49 +0000 (10:05 -0500)]
tests: specify the devices for OSDs in the scenario for collocation

Signed-off-by: Alfredo Deza <adeza@redhat.com>
Resolves: colocated#tests

8 years agotests: create tests for collocated journals
Alfredo Deza [Tue, 29 Nov 2016 15:05:16 +0000 (10:05 -0500)]
tests: create tests for collocated journals

Signed-off-by: Alfredo Deza <adeza@redhat.com>
Resolves: colocated#tests

8 years agocontrib: untrap before calling popd
Sébastien Han [Mon, 28 Nov 2016 22:22:49 +0000 (23:22 +0100)]
contrib: untrap before calling popd

we need to untrap before calling popd if the script is being run from a
directory different than the git repo. popd will go back to the previous
directory and trap will execute reset_hard_origin function on EXIT.
If we are not in the righ directory, this will fail.

Signed-off-by: Sébastien Han <seb@redhat.com>
8 years agoMerge pull request #1127 from ceph/osd-scenarios
Alfredo Deza [Mon, 28 Nov 2016 21:47:50 +0000 (16:47 -0500)]
Merge pull request #1127 from ceph/osd-scenarios

testing dmcrypt OSD scenarios

8 years agotests: change centos7-cluster to use raw_multi_journal 1127/head
Andrew Schoen [Mon, 28 Nov 2016 21:10:07 +0000 (15:10 -0600)]
tests: change centos7-cluster to use raw_multi_journal

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
8 years agotests: change xenial-cluster to use raw_multi_journal
Andrew Schoen [Mon, 28 Nov 2016 21:05:48 +0000 (15:05 -0600)]
tests: change xenial-cluster to use raw_multi_journal

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
8 years agotesting: change raw-multi-journal to journal-collocation
Andrew Schoen [Mon, 28 Nov 2016 21:02:16 +0000 (15:02 -0600)]
testing: change raw-multi-journal to journal-collocation

We want to test journal collocation here because we're gonna switch
xenial-cluster and centos7-cluster to use a dedicated journal.

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
8 years agotesting: rename the dmcrypt-dedicated-journal to dmcrypt-journal
Andrew Schoen [Mon, 28 Nov 2016 20:58:09 +0000 (14:58 -0600)]
testing: rename the dmcrypt-dedicated-journal to dmcrypt-journal

We're going to assume all scenarios use a dedicated journal unless
otherwise specified.

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
8 years agocontrib: don't fetch tags since we push force
Sébastien Han [Mon, 28 Nov 2016 18:07:52 +0000 (19:07 +0100)]
contrib: don't fetch tags since we push force

reverting 4be9c52bafdf4b55edde2aebce9607268a7871e3

Signed-off-by: Sébastien Han <seb@redhat.com>
8 years agotests: make Vagrantfile a symlink for dmcrypt scenarios
Andrew Schoen [Mon, 28 Nov 2016 18:06:25 +0000 (12:06 -0600)]
tests: make Vagrantfile a symlink for dmcrypt scenarios

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
8 years agotests: adds a scenario using OSD dmcrypt_journal_collocation
Andrew Schoen [Mon, 28 Nov 2016 18:03:00 +0000 (12:03 -0600)]
tests: adds a scenario using OSD dmcrypt_journal_collocation

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
8 years agotests: adds a dmcrypt-dedicated-journal testing scenario
Andrew Schoen [Mon, 28 Nov 2016 17:20:31 +0000 (11:20 -0600)]
tests: adds a dmcrypt-dedicated-journal testing scenario

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
8 years agocontrib: use push force because of the CI
Sébastien Han [Mon, 28 Nov 2016 17:54:14 +0000 (18:54 +0100)]
contrib: use push force because of the CI

If we don't force push we won't be able to push new content from
jenkins.

Signed-off-by: Sébastien Han <seb@redhat.com>