]> git.apps.os.sepia.ceph.com Git - ceph-ansible.git/log
ceph-ansible.git
8 years agoBetter --check compatibility for ceph-rgw role 1121/head
Daniel Marks [Sun, 27 Nov 2016 13:27:28 +0000 (14:27 +0100)]
Better --check compatibility for ceph-rgw role

Carefully chosen "always_run: true" parameters for read-only tasks that
register variables. This enables --check runs (at least on deployed
clusters).

8 years agoBetter --check compatibility for ceph-restapi role
Daniel Marks [Sun, 27 Nov 2016 13:16:36 +0000 (14:16 +0100)]
Better --check compatibility for ceph-restapi role

Carefully chosen "always_run: true" parameters for read-only tasks that
register variables. This enables --check runs (at least on deployed
clusters).

8 years agoBetter --check compatibility for ceph-rbd-mirror role
Daniel Marks [Sat, 26 Nov 2016 21:21:42 +0000 (22:21 +0100)]
Better --check compatibility for ceph-rbd-mirror role

Carefully chosen "always_run: true" parameters for read-only tasks that
register variables. This enables --check runs (at least on deployed
clusters).

8 years agoBetter --check compatibility for ceph-osd role
Daniel Marks [Sat, 26 Nov 2016 21:16:07 +0000 (22:16 +0100)]
Better --check compatibility for ceph-osd role

Carefully chosen "always_run: true" parameters for read-only tasks that
register variables. This enables --check runs (at least on deployed
clusters).

8 years agoBetter --check compatibility for ceph-nfs role
Daniel Marks [Sat, 26 Nov 2016 20:31:56 +0000 (21:31 +0100)]
Better --check compatibility for ceph-nfs role

Carefully chosen "always_run: true" parameters for read-only tasks that
register variables. This enables --check runs (at least on deployed
clusters).

8 years agoBetter --check compatibility for ceph-mon role
Daniel Marks [Sat, 26 Nov 2016 20:05:32 +0000 (21:05 +0100)]
Better --check compatibility for ceph-mon role

Carefully chosen "always_run: true" parameters for read-only tasks that
register variables. This enables --check runs (at least on deployed
clusters).

8 years agoBetter --check compatibility for ceph-mds role
Daniel Marks [Sat, 26 Nov 2016 19:52:24 +0000 (20:52 +0100)]
Better --check compatibility for ceph-mds role

Carefully chosen "always_run: true" parameters for read-only tasks that
register variables. This enables --check runs (at least on deployed
clusters).

8 years agoBetter --check compatibility for ceph-fetch-keys role
Daniel Marks [Sat, 26 Nov 2016 19:42:08 +0000 (20:42 +0100)]
Better --check compatibility for ceph-fetch-keys role

Carefully chosen "always_run: true" parameters for read-only tasks that
register variables. This enables --check runs (at least on deployed
clusters).

8 years agoBetter --check compatibility for ceph-common role
Daniel Marks [Sat, 26 Nov 2016 19:39:12 +0000 (20:39 +0100)]
Better --check compatibility for ceph-common role

Carefully chosen "always_run: true" parameters for read-only tasks that
register variables. This enables --check runs (at least on deployed
clusters).

8 years agoBetter --check compatibility for ceph-common-coreos role
Daniel Marks [Sat, 26 Nov 2016 16:48:06 +0000 (17:48 +0100)]
Better --check compatibility for ceph-common-coreos role

Carefully chosen "always_run: true" parameters for read-only tasks that
register variables. This enables --check runs (at least on deployed
clusters).

8 years agoBetter --check compatibility for ceph-agent role
Daniel Marks [Sat, 26 Nov 2016 16:35:57 +0000 (17:35 +0100)]
Better --check compatibility for ceph-agent role

Carefully chosen "always_run: true" parameters for read-only tasks that
register variables. This enables --check runs (at least on deployed
clusters).

8 years agoMerge pull request #1102 from Logan2211/fix-default-syntax
Sébastien Han [Fri, 25 Nov 2016 15:07:34 +0000 (16:07 +0100)]
Merge pull request #1102 from Logan2211/fix-default-syntax

Fix jinja defaulting syntax

8 years agoMerge pull request #1118 from guits/master
Sébastien Han [Fri, 25 Nov 2016 10:21:51 +0000 (11:21 +0100)]
Merge pull request #1118 from guits/master

[shrink-mon]: force playbook to fail if there is only one mon

8 years ago[shrink-mon]: force playbook to fail if there is only one mon 1118/head
Guillaume Abrioux [Thu, 24 Nov 2016 21:46:46 +0000 (22:46 +0100)]
[shrink-mon]: force playbook to fail if there is only one mon

The playbook will fail if only 1 mon is in the cluster
and advise to use the `purge-cluster` playbook instead.

Fix #1083

8 years agoFix jinja defaulting syntax 1102/head
Logan V [Thu, 17 Nov 2016 17:10:01 +0000 (11:10 -0600)]
Fix jinja defaulting syntax

a397922 introduced a syntax error by attempting to default an unquoted
string, which causes execution failures on some ansible versions with:

Failed to template {{ ceph_rhcs_mount_path }}: Failed to template {{ ceph_stable_rh_storage_mount_path | default(/tmp/rh-storage-mount) }}: template error while templating string: unexpected '/'

8 years agoMerge pull request #1117 from D3N14L/master
Sébastien Han [Thu, 24 Nov 2016 15:43:15 +0000 (16:43 +0100)]
Merge pull request #1117 from D3N14L/master

Improve --check compatibility for ceph-common, ceph-mon and ceph-rgw roles

8 years agoMerge pull request #1113 from guits/master
Sébastien Han [Thu, 24 Nov 2016 14:19:29 +0000 (15:19 +0100)]
Merge pull request #1113 from guits/master

All `include_vars` need to have `*.yml`, `*.yaml` or `*.json` extension.

8 years agoAll `include_vars` need to have `*.yml`, `*.yaml` or `*.json` extension. 1113/head
Guillaume Abrioux [Wed, 23 Nov 2016 10:14:32 +0000 (11:14 +0100)]
All `include_vars` need to have `*.yml`, `*.yaml` or `*.json` extension.

As introduced in the following PR:
- https://github.com/ansible/ansible/pull/17207
we need to refactor our code.

8 years agoBetter --check compatibility for ceph-rgw role 1117/head
Daniel Marks [Thu, 24 Nov 2016 10:54:43 +0000 (11:54 +0100)]
Better --check compatibility for ceph-rgw role

Carefully chosen "always_run: true" parameters for read-only tasks that
register variables. This enables --check runs (at least on deployed
clusters).

8 years agoBetter --check compatibility for ceph-mon role
Daniel Marks [Thu, 24 Nov 2016 10:54:10 +0000 (11:54 +0100)]
Better --check compatibility for ceph-mon role

Carefully chosen "always_run: true" parameters for read-only tasks that
register variables. This enables --check runs (at least on deployed
clusters).

8 years agoBetter --check compatibility for ceph-common role
Daniel Marks [Thu, 24 Nov 2016 10:53:29 +0000 (11:53 +0100)]
Better --check compatibility for ceph-common role

Carefully chosen "always_run: true" parameters for read-only tasks that
register variables. This enables --check runs (at least on deployed
clusters).

8 years agocontrib: add a new script to update remote repos
Sébastien Han [Mon, 21 Nov 2016 17:31:10 +0000 (18:31 +0100)]
contrib: add a new script to update remote repos

This script allows us to split ceph-ansible role directory into
repositories that can be pushed.
This script is an enhanced version of splitup.yml that was using git
subtree. I haven't found any straighforward way to push tag using git
subtree where git filter-branch seems easier and more practical.

Now we can not only push commits from master but also all the relative
tags. It is nice for projects that want to consume our roles from the
Ansible Galaxy, pointing to a specific tags looking for a particular
release.

For backward compatobility, I pushed a branch called
"master-history-bkp" on all the github.com/ceph/ansbible-ceph-*
repositories. So current projects using a SHA to retrieve the roles
won't be impacted by this change.

At the time of the commit, running this script took  104.60s to
complete.

Signed-off-by: Sébastien Han <seb@redhat.com>
Co-Authored-By: Erwan Velu <erwan@redhat.com>
Co-Authored-By: Logan Vig <logan2211@gmail.com>
8 years agocontrib: add a new script to update remote repos
Sébastien Han [Mon, 21 Nov 2016 17:31:10 +0000 (18:31 +0100)]
contrib: add a new script to update remote repos

This script allows us to split ceph-ansible role directory into
repositories that can be pushed.
This script is an enhanced version of splitup.yml that was using git
subtree. I haven't found any straighforward way to push tag using git
subtree where git filter-branch seems easier and more practical.

Now we can not only push commits from master but also all the relative
tags. It is nice for projects that want to consume our roles from the
Ansible Galaxy, pointing to a specific tags looking for a particular
release.

For backward compatobility, I pushed a branch called
"master-history-bkp" on all the github.com/ceph/ansbible-ceph-*
repositories. So current projects using a SHA to retrieve the roles
won't be impacted by this change.

At the time of the commit, running this script took  104.60s to
complete.

Signed-off-by: Sébastien Han <seb@redhat.com>
Co-Authored-By: Erwan Velu <erwan@redhat.com>
Co-Authored-By: Logan Vig <logan2211@gmail.com>
8 years agoMake NTP checks usable with '--check' 1115/head
Daniel Marks [Wed, 23 Nov 2016 12:52:34 +0000 (13:52 +0100)]
Make NTP checks usable with '--check'

As those task are only reading the installed packages and their output
is required in later tasks it is safe to run them also in check mode.

8 years agoMerge pull request #1112 from zhsj/fix-py3
Sébastien Han [Tue, 22 Nov 2016 17:07:09 +0000 (18:07 +0100)]
Merge pull request #1112 from zhsj/fix-py3

fix python3 compatibility in jinja2 template.

8 years agoMerge pull request #1101 from ceph/ci-testing
Alfredo Deza [Tue, 22 Nov 2016 16:41:05 +0000 (11:41 -0500)]
Merge pull request #1101 from ceph/ci-testing

testing: changes to allow for concurrent functional tests

8 years agoMerge pull request #1077 from font/rolling_update
Sébastien Han [Tue, 22 Nov 2016 15:56:46 +0000 (16:56 +0100)]
Merge pull request #1077 from font/rolling_update

Support containerized rolling update

8 years agoMerge pull request #1109 from ceph/bz-1396742
Sébastien Han [Tue, 22 Nov 2016 15:10:17 +0000 (16:10 +0100)]
Merge pull request #1109 from ceph/bz-1396742

rolling_update: clarify "serial" usage

8 years agofix python3 compatibility in jinja2 template. 1112/head
Shengjing Zhu [Tue, 22 Nov 2016 13:52:52 +0000 (21:52 +0800)]
fix python3 compatibility in jinja2 template.

rename iterkeys to keys

8 years agotesting: each scenario should use it's own fetch_directory 1101/head
Andrew Schoen [Mon, 21 Nov 2016 23:08:39 +0000 (17:08 -0600)]
testing: each scenario should use it's own fetch_directory

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
8 years agorolling_update: clarify "serial" usage 1109/head
Sébastien Han [Mon, 21 Nov 2016 10:29:05 +0000 (11:29 +0100)]
rolling_update: clarify "serial" usage

Prior to this commit the serial variable was poorly documented. Now we
are making clear that this value should be left untouched as the rolling
update mechanism should happen serially.

Solves: bz-1396742
Signed-off-by: Sébastien Han <seb@redhat.com>
8 years agoMerge pull request #1104 from ktdreyer/rm-apache
Sébastien Han [Mon, 21 Nov 2016 10:13:19 +0000 (11:13 +0100)]
Merge pull request #1104 from ktdreyer/rm-apache

remove apache support for RGW

8 years agotesting: remove the timeout for the ansible-playbook call
Andrew Schoen [Fri, 18 Nov 2016 19:21:29 +0000 (13:21 -0600)]
testing: remove the timeout for the ansible-playbook call

This isn't available on OSX and we'll end up hitting this timeout when
we don't want to, let's just remove it.

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
8 years agotests: update the subnet in the ceph.conf tests
Andrew Schoen [Fri, 18 Nov 2016 19:19:03 +0000 (13:19 -0600)]
tests: update the subnet in the ceph.conf tests

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
8 years agotests: add a UUID to the disk name when using libvirt
Andrew Schoen [Fri, 18 Nov 2016 17:09:20 +0000 (11:09 -0600)]
tests: add a UUID to the disk name when using libvirt

This will allow each testing scenario to have a unique names
for it's disks so there will not be conflicts when running tests
in parallel.

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
8 years agoremove apache support for RGW 1104/head ktdreyer/rm-apache
Ken Dreyer [Thu, 17 Nov 2016 22:43:31 +0000 (15:43 -0700)]
remove apache support for RGW

libfcgi is dead upstream (http://tracker.ceph.com/issues/16784)

The RGW developers intend to remove libfcgi support entirely before the
Luminous release.

Since libfcgi gets little-to-no developer attention or testing, remove
it entirely from ceph-ansible.

8 years agotests: fix monitor_address in ubuntu/16.04/mon/hosts
Andrew Schoen [Fri, 18 Nov 2016 17:02:47 +0000 (11:02 -0600)]
tests: fix monitor_address in ubuntu/16.04/mon/hosts

We changed the subnet in vagrant_variables.yml so this needs to change
here as well.

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
8 years agotesting: use the same sequence in the group_vars for every test scenario
Alfredo Deza [Fri, 18 Nov 2016 16:47:38 +0000 (11:47 -0500)]
testing: use the same sequence in the group_vars for every test scenario

Signed-off-by: Alfredo Deza <adeza@redhat.com>
8 years agotesting: use a sequence in the subnets for every vagrant_variables.yml to avoid colli...
Alfredo Deza [Fri, 18 Nov 2016 16:47:06 +0000 (11:47 -0500)]
testing: use a sequence in the subnets for every vagrant_variables.yml to avoid collisions

Signed-off-by: Alfredo Deza <adeza@redhat.com>
8 years agotesting: vagranfile should use random hostnames to prevent collisions
Alfredo Deza [Fri, 18 Nov 2016 16:44:09 +0000 (11:44 -0500)]
testing: vagranfile should use random hostnames to prevent collisions

Signed-off-by: Alfredo Deza <adeza@redhat.com>
8 years agoMerge pull request #1107 from font/osd_keyring_err v2.0.0
Sébastien Han [Fri, 18 Nov 2016 14:18:01 +0000 (15:18 +0100)]
Merge pull request #1107 from font/osd_keyring_err

Fix error caused by missing OSD cluster keyring

8 years agoMerge pull request #1106 from ktdreyer/rpm-packaging
Sébastien Han [Fri, 18 Nov 2016 14:16:14 +0000 (15:16 +0100)]
Merge pull request #1106 from ktdreyer/rpm-packaging

add RPM packaging for CI

8 years agoFix error caused by missing OSD cluster keyring 1107/head
Ivan Font [Fri, 18 Nov 2016 07:03:11 +0000 (23:03 -0800)]
Fix error caused by missing OSD cluster keyring

Ansible task was not properly fetching OSD cluster keyring causing
the keyring to be missing when we needed to authenticate. Similarly, we
were not properly waiting on the OSD keyring to be available before
continuing.

Signed-off-by: Ivan Font <ifont@redhat.com>
8 years agoadd RPM packaging for CI 1106/head
Ken Dreyer [Thu, 17 Nov 2016 22:21:20 +0000 (15:21 -0700)]
add RPM packaging for CI

"make rpm" will build a ceph-ansible RPM and place it in the current
working directory.

This will allow us to run this command in Jenkins for every branch.

8 years agoMerge pull request #1105 from ceph/libvirt-domain-conflicts
Alfredo Deza [Thu, 17 Nov 2016 21:27:33 +0000 (16:27 -0500)]
Merge pull request #1105 from ceph/libvirt-domain-conflicts

tests: sets random_hostname to true for libvirt nodes

8 years agotests: sets random_hostname to true for libvirt nodes 1105/head
Andrew Schoen [Thu, 17 Nov 2016 21:25:06 +0000 (15:25 -0600)]
tests: sets random_hostname to true for libvirt nodes

This will keep us from having name conflicts between our different
testing scenarios.

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
8 years agoRolling update changes for containerized deployments 1077/head
Ivan Font [Wed, 16 Nov 2016 08:35:09 +0000 (00:35 -0800)]
Rolling update changes for containerized deployments

Separate out systemd restart tasks for containerized and
non-containerized deployments

Signed-off-by: Ivan Font <ifont@redhat.com>
8 years agoWarn user when upgrading cluster with only one mon
Ivan Font [Mon, 14 Nov 2016 19:04:43 +0000 (11:04 -0800)]
Warn user when upgrading cluster with only one mon

Signed-off-by: Ivan Font <ifont@redhat.com>
8 years agoSupport containerized rolling update
Ivan Font [Sun, 6 Nov 2016 04:15:26 +0000 (21:15 -0700)]
Support containerized rolling update

- Update rolling update playbook to support containerized deployments
  for mons, osds, mdss, and rgws
- Skip checking if existing cluster is running when performing a rolling
  update
- Fixed bug where we were failing to start the mds container because it
  was missing the admin keyring. The admin keyring was missing because
  it was not being pushed from the mon host to the ansible host due to
  the keyring not being available before running the copy_configs.yml
  task include file. Now we forcefully wait for the admin keyring to be
  generated before continuing with the copy_configs.yml task include file
- Skip pre_requisite.yml when running on atomic host. This technically
  no longer requires specifying to skip tasks containing the with_pkg tag
- Add missing variables to all.docker.sample
- Misc. cleanup

Signed-off-by: Ivan Font <ifont@redhat.com>
8 years agoMerge pull request #1103 from ceph/gatherkeys-timeout
Andrew Schoen [Thu, 17 Nov 2016 17:49:55 +0000 (11:49 -0600)]
Merge pull request #1103 from ceph/gatherkeys-timeout

testing: add a timeout per ansible call to workaround possible gatherkeys issues

8 years agotesting: add a timeout per ansible call to workaround possible gatherkeys issues 1103/head
Alfredo Deza [Thu, 17 Nov 2016 17:47:49 +0000 (12:47 -0500)]
testing: add a timeout per ansible call to workaround possible gatherkeys issues

Signed-off-by: Alfredo Deza <adeza@redhat.com>
8 years agoMerge pull request #1099 from jsaintrocc/checkmode-fix
Sébastien Han [Thu, 17 Nov 2016 16:02:22 +0000 (17:02 +0100)]
Merge pull request #1099 from jsaintrocc/checkmode-fix

Replaced Ansible 2.2 specific `check_mode: no`

8 years agoMerge pull request #1100 from ceph/rhbz-1395073
Sébastien Han [Thu, 17 Nov 2016 15:49:08 +0000 (16:49 +0100)]
Merge pull request #1100 from ceph/rhbz-1395073

rolling_update: bump retries for osd_check/retries to 20 minutes

8 years agorolling_update: bump retries for osd_check/retries to 20 minutes 1100/head
Alfredo Deza [Thu, 17 Nov 2016 15:43:58 +0000 (10:43 -0500)]
rolling_update: bump retries for osd_check/retries to 20 minutes

Signed-off-by: Alfredo Deza <adeza@redhat.com>
Resolves: rhbz#1395073

8 years agoReplaced Ansible 2.2 specific check_mode: no with backwards compatible always_run... 1099/head
James Saint-Rossy [Wed, 16 Nov 2016 21:47:43 +0000 (16:47 -0500)]
Replaced Ansible 2.2 specific check_mode: no with backwards compatible always_run: yes

8 years agoMerge pull request #1098 from ceph/libvirt-devices
Alfredo Deza [Wed, 16 Nov 2016 21:20:35 +0000 (16:20 -0500)]
Merge pull request #1098 from ceph/libvirt-devices

tests: make libvirt use the same device names as virtualbox

8 years agotests: make libvirt use the same device names as virtualbox 1098/head
Andrew Schoen [Wed, 16 Nov 2016 19:01:22 +0000 (13:01 -0600)]
tests: make libvirt use the same device names as virtualbox

This makes our libvirt boxes come up with the OS on /dev/vda and
three devices added at /dev/sd{a/b/c} so that we can ensure that
the OSD devices we want to use can always be available for
both virtualbox and libvirt for both xenial and centos7.

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
8 years agoMerge pull request #1097 from ceph/ansible-ci-version v2.0
Sébastien Han [Wed, 16 Nov 2016 15:46:59 +0000 (16:46 +0100)]
Merge pull request #1097 from ceph/ansible-ci-version

ci: bump ansible version to 2.2

8 years agoci: bump ansible version to 2.2 1097/head
Sébastien Han [Wed, 16 Nov 2016 15:44:45 +0000 (16:44 +0100)]
ci: bump ansible version to 2.2

ceph-ansible is now fully compliant with ansible 2.2 so bumping the
ansible version to make we test the right things.

Signed-off-by: Sébastien Han <seb@redhat.com>
8 years agoMerge pull request #1092 from ceph/ceph-boxes
Andrew Schoen [Wed, 16 Nov 2016 15:42:31 +0000 (09:42 -0600)]
Merge pull request #1092 from ceph/ceph-boxes

Use new ceph vagrant boxes

8 years agoMerge pull request #1068 from ceph/v2.2
Sébastien Han [Wed, 16 Nov 2016 15:33:40 +0000 (16:33 +0100)]
Merge pull request #1068 from ceph/v2.2

moving to ansible v2.2 compatibility

8 years agoMerge pull request #1082 from albertomurillo/v2.2 1068/head
Sébastien Han [Wed, 9 Nov 2016 08:53:22 +0000 (09:53 +0100)]
Merge pull request #1082 from albertomurillo/v2.2

sync config_template plugin with master branch

8 years agoMerge pull request #1096 from ceph/cleanup-yml-tests
Sébastien Han [Wed, 16 Nov 2016 13:52:06 +0000 (14:52 +0100)]
Merge pull request #1096 from ceph/cleanup-yml-tests

tests: remove all validating yml playbooks

8 years agotests: remove all validating yml playbooks 1096/head
Alfredo Deza [Wed, 16 Nov 2016 13:49:51 +0000 (08:49 -0500)]
tests: remove all validating yml playbooks

Signed-off-by: Alfredo Deza <adeza@redhat.com>
8 years agotesting: define the ceph box in the sample vagrant_variable file 1092/head
Alfredo Deza [Tue, 15 Nov 2016 12:51:47 +0000 (07:51 -0500)]
testing: define the ceph box in the sample vagrant_variable file

Signed-off-by: Alfredo Deza <adeza@redhat.com>
8 years agotesting: disable vagrant sync folder add on xenial mon osd scenario
Alfredo Deza [Mon, 14 Nov 2016 19:45:57 +0000 (14:45 -0500)]
testing: disable vagrant sync folder add on xenial mon osd scenario

Signed-off-by: Alfredo Deza <adeza@redhat.com>
8 years agotesting: disable vagrant sync folder add on xenial mon osd mds rgw scenario
Alfredo Deza [Mon, 14 Nov 2016 19:45:43 +0000 (14:45 -0500)]
testing: disable vagrant sync folder add on xenial mon osd mds rgw scenario

Signed-off-by: Alfredo Deza <adeza@redhat.com>
8 years agotesting: disable vagrant sync folder add on centos 7 mon osd mds rgw scenario
Alfredo Deza [Mon, 14 Nov 2016 19:45:07 +0000 (14:45 -0500)]
testing: disable vagrant sync folder add on centos 7 mon osd mds rgw scenario

Signed-off-by: Alfredo Deza <adeza@redhat.com>
8 years agotesting: disable vagrant sync folder add on centos 7 mon osd scenario
Alfredo Deza [Mon, 14 Nov 2016 19:44:49 +0000 (14:44 -0500)]
testing: disable vagrant sync folder add on centos 7 mon osd scenario

Signed-off-by: Alfredo Deza <adeza@redhat.com>
8 years agotesting: use ceph vagrant box on ubuntu xenial mon osd scenario
Alfredo Deza [Mon, 14 Nov 2016 19:43:02 +0000 (14:43 -0500)]
testing: use ceph vagrant box on ubuntu xenial mon osd scenario

Signed-off-by: Alfredo Deza <adeza@redhat.com>
8 years agotesting: use ceph vagrant box on ubuntu xenial mon osd mds rgw scenario
Alfredo Deza [Mon, 14 Nov 2016 19:42:35 +0000 (14:42 -0500)]
testing: use ceph vagrant box on ubuntu xenial mon osd mds rgw scenario

Signed-off-by: Alfredo Deza <adeza@redhat.com>
8 years agotesting: add comment for ceph vagrant boxes on centos 7 mon osd scenario
Alfredo Deza [Mon, 14 Nov 2016 19:41:39 +0000 (14:41 -0500)]
testing: add comment for ceph vagrant boxes on centos 7 mon osd scenario

Signed-off-by: Alfredo Deza <adeza@redhat.com>
8 years agotesting: add comment for ceph vagrant boxes on centos 7 mon osd mds rgw scenario
Alfredo Deza [Mon, 14 Nov 2016 19:41:23 +0000 (14:41 -0500)]
testing: add comment for ceph vagrant boxes on centos 7 mon osd mds rgw scenario

Signed-off-by: Alfredo Deza <adeza@redhat.com>
8 years agotesting: parse and set DISABLE_SYNCED_FOLDER in vagrant file
Alfredo Deza [Mon, 14 Nov 2016 18:25:26 +0000 (13:25 -0500)]
testing: parse and set DISABLE_SYNCED_FOLDER in vagrant file

Signed-off-by: Alfredo Deza <adeza@redhat.com>
8 years agotests: disable sync folder when testing
Alfredo Deza [Mon, 14 Nov 2016 18:24:25 +0000 (13:24 -0500)]
tests: disable sync folder when testing

Signed-off-by: Alfredo Deza <adeza@redhat.com>
8 years agoMerge pull request #1093 from ceph/RHBZ-1394929
Alfredo Deza [Mon, 14 Nov 2016 21:10:06 +0000 (16:10 -0500)]
Merge pull request #1093 from ceph/RHBZ-1394929

rolling update: stop services before upgrading and start afterwards

8 years agorolling update: stop RGWs before upgrade and start afterwards 1093/head
Andrew Schoen [Mon, 14 Nov 2016 20:44:55 +0000 (14:44 -0600)]
rolling update: stop RGWs before upgrade and start afterwards

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
Resolves: rhbz#1394929

8 years agorolling update: stop MDSs before upgrading and start afterwards
Andrew Schoen [Mon, 14 Nov 2016 20:42:52 +0000 (14:42 -0600)]
rolling update: stop MDSs before upgrading and start afterwards

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
Resolves: rhbz#1394929

8 years agorolling update: stop MONs before upgrading and start afterwards
Andrew Schoen [Mon, 14 Nov 2016 20:40:31 +0000 (14:40 -0600)]
rolling update: stop MONs before upgrading and start afterwards

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
Resolves: rhbz#1394929

8 years agorolling update: stop OSDs before upgrading
Andrew Schoen [Mon, 14 Nov 2016 20:25:46 +0000 (14:25 -0600)]
rolling update: stop OSDs before upgrading

This avoids a bug where OSDs are sometimes restarted twice on
upgrades which leaves the OSD process running but not marked up.

See:

https://bugzilla.redhat.com/show_bug.cgi?id=1394928
https://bugzilla.redhat.com/show_bug.cgi?id=1391675
https://bugzilla.redhat.com/show_bug.cgi?id=1394929

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
Resolves: rhbz#1394929

8 years agoMerge pull request #1091 from ceph/rgw-when
Sébastien Han [Mon, 14 Nov 2016 15:25:26 +0000 (16:25 +0100)]
Merge pull request #1091 from ceph/rgw-when

ceph-rgw: add systemd condition

8 years agoMerge pull request #1032 from ceph/backward-compat-rhcs
Sébastien Han [Mon, 14 Nov 2016 14:49:42 +0000 (15:49 +0100)]
Merge pull request #1032 from ceph/backward-compat-rhcs

common: backward compatibility name for rhcs

8 years agoceph-rgw: add systemd condition 1091/head
Sébastien Han [Mon, 14 Nov 2016 14:12:20 +0000 (15:12 +0100)]
ceph-rgw: add systemd condition

We have a fact that detects the package manager, so we can detect if
systemd is used. Radosgw was still using some old logic from Ubuntu.
Ubuntu 16.04 now has systemd so we don't need to configure rgw as it was
running on upstart.

Signed-off-by: Sébastien Han <seb@redhat.com>
8 years agoMerge pull request #1089 from ceph/vagrant-boxes
Alfredo Deza [Thu, 10 Nov 2016 21:16:07 +0000 (16:16 -0500)]
Merge pull request #1089 from ceph/vagrant-boxes

tests: adds a few more scenarios

8 years agotests: adds comments to tox.ini describing each scenario 1089/head
Andrew Schoen [Thu, 10 Nov 2016 18:04:56 +0000 (12:04 -0600)]
tests: adds comments to tox.ini describing each scenario

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
Resolves: vagrant#boxes

8 years agotests: make the centos7-mon-osd scenario use raw_multi_journal
Andrew Schoen [Thu, 10 Nov 2016 17:59:45 +0000 (11:59 -0600)]
tests: make the centos7-mon-osd scenario use raw_multi_journal

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
Resolves: vagrant#boxes

8 years agotests: adds the centos7-cluster scenario to tox.ini
Andrew Schoen [Wed, 9 Nov 2016 20:52:37 +0000 (14:52 -0600)]
tests: adds the centos7-cluster scenario to tox.ini

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
Resolves: vagrant#boxes

8 years agotests: adds a scenario for a centos7 1mon 1osd 1mds 1rgw cluster
Andrew Schoen [Wed, 9 Nov 2016 20:51:54 +0000 (14:51 -0600)]
tests: adds a scenario for a centos7 1mon 1osd 1mds 1rgw cluster

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
Resolves: vagrant#boxes

8 years agotests: adds the xenial mon-osd-mds-rgw scenario as xenial-cluster
Andrew Schoen [Wed, 9 Nov 2016 20:16:12 +0000 (14:16 -0600)]
tests: adds the xenial mon-osd-mds-rgw scenario as xenial-cluster

I tried adding the scenario as xenial-mon-osd-mds-rgw but for
whatever reason tox refused to select that environment. Changing the
name to xenial-cluster fixed that though.

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
Resolves: vagrant#boxes

8 years agotests: adds a scenario to test a xenial 1mon 1osd 1mds 1rgw cluster
Andrew Schoen [Wed, 9 Nov 2016 20:14:10 +0000 (14:14 -0600)]
tests: adds a scenario to test a xenial 1mon 1osd 1mds 1rgw cluster

This scenario duplicates what we are currently doing with our
ceph-ansible testing using OVH and a single node, except now
we are using 4 separate nodes.

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
Resolves: vagrant#boxes

8 years agotests: make xenial-mon-osd use raw_multi_journal instead of collocation
Andrew Schoen [Wed, 9 Nov 2016 20:12:59 +0000 (14:12 -0600)]
tests: make xenial-mon-osd use raw_multi_journal instead of collocation

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
Resolves: vagrant#boxes

8 years agoMerge pull request #1087 from ceph/vagrant-boxes
Sébastien Han [Wed, 9 Nov 2016 11:09:00 +0000 (12:09 +0100)]
Merge pull request #1087 from ceph/vagrant-boxes

use vagrant boxes that support libvirt and virtualbox

8 years agotests: provide a way use libvirt with vagrant and tox 1087/head
Andrew Schoen [Tue, 8 Nov 2016 23:00:09 +0000 (17:00 -0600)]
tests: provide a way use libvirt with vagrant and tox

This allows for:

tox -- --provider=libvirt

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
Resolves: vagrant#boxes

8 years agotests: change the centos box to centos/7
Andrew Schoen [Tue, 8 Nov 2016 21:53:58 +0000 (15:53 -0600)]
tests: change the centos box to centos/7

This box supports both libvirt and virtualbox. Eventually we want to
be building our own boxes but this should work in the short term.

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
8 years agotests: change the xenial box to yk0/ubuntu-xenial
Andrew Schoen [Tue, 8 Nov 2016 21:48:16 +0000 (15:48 -0600)]
tests: change the xenial box to yk0/ubuntu-xenial

This box supports both virtualbox and libvirt. Eventually
we want to be building our own vagrant boxes, but this might
work for now.

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
Resolves: vagrant#boxes

8 years agoMerge pull request #1086 from ceph/testing-updates
Alfredo Deza [Tue, 8 Nov 2016 18:39:32 +0000 (13:39 -0500)]
Merge pull request #1086 from ceph/testing-updates

testing: adds 3 new functional testing scenarios

8 years agotests: adds the vagrant_variables.yml that were previously ignored 1086/head
Andrew Schoen [Tue, 8 Nov 2016 17:29:47 +0000 (11:29 -0600)]
tests: adds the vagrant_variables.yml that were previously ignored

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
Resolves: testing#updates

8 years agodon't ignore vagrant_variables.yml in test directories
Andrew Schoen [Tue, 8 Nov 2016 17:03:25 +0000 (11:03 -0600)]
don't ignore vagrant_variables.yml in test directories

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
Resolves: testing#updates

8 years agotests: updates tox to test the new scenarios
Andrew Schoen [Tue, 8 Nov 2016 16:34:16 +0000 (10:34 -0600)]
tests: updates tox to test the new scenarios

This also turns on -vv for ansible and adds an environment for
ansible==1.9.4 testing.

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
Resolves: testing#updates

8 years agotests: adds a 1mon 1osd centos7 test scenario
Andrew Schoen [Tue, 8 Nov 2016 16:32:39 +0000 (10:32 -0600)]
tests: adds a 1mon 1osd centos7 test scenario

This uses journal collocation for the osd.

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
Resolves: testing#updates