This enables ceph volume backup driver in scenario 001 so that we can
test deployment of the driver.
Change-Id: Ia07f1be0491bb25a0e5d7b518e9c764a6281474d
This enables manila with cephfs backend in scenario 004 test.
Depends-on: https://review.opendev.org/884461
Change-Id: I60d6ce301cf714cd918becaf493eb7396ec9631a
to avoid warning caused by too many pgs assigned to a single OSD.
The warning is seen in scenario 004 integration job because Rados
Gateway creates additional pools.
Change-Id: Ice01e0ea8313ae9732c0e48003f886fb85ecc9fd
The previous setting has been using the wrong option and did not limit
the host which rados gw listens to.
Change-Id: I0f51d56df8a73190071dd9e7568fcf3870ec1e8d
civetweb was deprecated since Ceph Pacific in favor of beast which has
been supported since Mimic.
Depends-on: https://review.opendev.org/853852
Change-Id: I278584d4e9be59e2fece94a9d135b54f331a5256
This is a partial revert of 0c938590d2bc7e56a04b196ee8cf835786b9a6d4
and removes workaround for old ceph package. Now python3-ceph should
be installed instead of python2-ceph.
Change-Id: Ib139a7df91cadbf9ac77eb8d94791e5865414ba9
Currently we support usage of distro packages only, and this custom
fact can be simply replaced by the default fact.
Change-Id: I8cad680e0ec96388cf5fa0e34f50b2c4d5969523
In Puppet 3 there was a need to use absolute
names to prevent issues. Since Puppet 4 everything
is absolute by default which makes this not needed.
We need to change this everywhere so that we can
revert the pin in [1] that now prevents us from
using the latest version of the puppet-lint check
puppet-lint-absolute_classname_check that properly
checks that we dont use absolute names.
[1] https://review.opendev.org/#/c/697742/
Change-Id: I78b74fbeb08be7234189e4d266412fb7fb7a73c0
[1] added lvm_create task to create a lvm device
required by ceph-nautilus. lvcreate command fails
some times with below Error:-
/dev/ceph_vg/lv_data: not found: device not cleared
To overcome this let's add a direct dependency on 'lvm2'
package as lvm_create task needs that package, in addition
adds a sleep task of 5 seconds.
[1] https://review.opendev.org/#/c/650281
Change-Id: I61b6c353cb5216265bf31d56a9f40437821b92fc
Ceph Nautilus has now a stable release 14.2.0 and it's being packaged
in CentOS Storage SIG. Mimic release was not LTS so let's replace it
by Nautilus where possible.
Since nautilus, the supported store backend is LVM which can use a block
device directly (not loopback) or a LVM logical volume for OSDs.
This patch is creating a LVM structure to create the OSD used to
validate Ceph and switching scenario004 to use nautilus instead of
mimic, so we will keep testing both Luminous and Nautilus, LTS releases.
Depends-On: https://review.opendev.org/#/c/650263/
Change-Id: I81441237a8b7cf651fa6926b2eb3c4ed0d42bb58
Disable SSL as currently there are known issues in services
to work with python3 + SSL on Fedora.
Also fixed following:-
- watcher to work without SSL
- Handle source /etc/bashrc for users don't have .bashrc created
- Install python3-rbd in Fedora and RedHat > 7
Change-Id: I5b67f40ec3c687f5282c65e10b4a1ee1fe5528f3
keystone is creating default roles for member since [1]. This
is "member" and it's not configurable so we need to switch it
in p-o-i too to avoid conflicting names errors (keystone is not
case sensitive in object names).
[1] https://review.openstack.org/#/c/572243/
Depends-On: Icd91bd27d0c0f375d2e99e9c018d3971fc49ce18
Change-Id: Id838d69e4ace9b2ea6c2a9cc6f8fd079d2115dd6
The Pike release of OpenStack now comes bundled with Ceph 12.x
which requires ceph-mgr in order to run properly. This patch
installs it.
The ceph-mgr service is not available in 10.x which is what
CentOS ships, we can drop the conditional once CentOS ships
the latest stable Ceph release.
Change-Id: I9b9c63c835770f7632051ac114f8bf6b19e8e210
The rbd_default_features extra configs parameter was removed from
this module as per I3242cfe0f141fe2bbb951ce2cf7071fd07456db8 so that,
as per I95cafec9160f5c3b1610a0408f2f80d567fcaa89, puppet-ceph could
manage it instead. This change passes the new rbd_default_features
parameter to puppet-ceph with a value of 15, the same value that this
module used to use for performance reasons.
Change-Id: I1810b0e58c47f3827b6396c039eb9a32510a7993
Depends-On: I477640c5db2cc84622da7734fb54cfe7740bd137
Remove the client/rbd_default_features resource from this module
so that the puppet-ceph module can manage it instead as per
I477640c5db2cc84622da7734fb54cfe7740bd137. This change was
anticpated as per I95cafec9160f5c3b1610a0408f2f80d567fcaa89.
Change-Id: I3242cfe0f141fe2bbb951ce2cf7071fd07456db8
Related-Bug: 1633552
puppet ceph used hardcoded values in ceph::rgw::keystone::auth
This updates p-o-i scenario004 to provide values to the previously
hardcoded parameters.
Change-Id: Ie9b8113351300ab8ab331f083a47fef7d4735e69
Introduce scenario004 that will test Ceph RGW with OpenStack.
Co-Authored-by: Emilien Macchi <emilien@redhat.com>
Co-Authored-by: Keith Schincke <keith.schincke@gmail.com>
Co-Authored-by: Andrew Woodward <awoodward@mirantis.com>
Change-Id: I0dc1ad61a5f7e4f149e368055f045846ce88c21c
We are working to land I477640c5db2cc84622da7734fb54cfe7740bd137
which creates a duplicate declration, we should be able to move from create
to ensure resource to minimize this until it lands.
Related_bug: 1633552
Change-Id: I95cafec9160f5c3b1610a0408f2f80d567fcaa89
This patch updates openstack_integration::ceph to use the
osd_max_object_name_len and osd_max_object_namespace_len parameters.
This will allow p-o-i jobs to use EXT4 file systems without
adding parameters via ceph_extra_config
Change-Id: Iea0dd5ae7bb1c438a68db0fd2af63bfa398efcd3
Depends-On: Id487d6d656e88dfe29eac80c378058551c92dfd9
Using /srv/data caused SElinux alerts:
https://bugzilla.redhat.com/show_bug.cgi?id=1354488
This patch is an attempt to reduce or remove the alerts in our CI and
use /var/lib/ceph directory to store OSD data, like recommended by Ceph
documentation.
Change-Id: I4e6f31838ffb2b9dcde73fc06fbff78cd084b0af
This patch will change the way we deployed Ceph on CentOS nodes.
* Use CentOS SIG by default
https://wiki.centos.org/SpecialInterestGroup/Storage/
* Stop deploying EPEL, which is conflicting with RDO packaging.
All our CI needs should be in RDO and SIG. Let's stop deploying EPEL at
all.
Depends-On: Ic42c19b2b5b4bc499565c69c3053395173c29a8e
Change-Id: I8d435a77c41c38354b20a40ab30dbc595cb6412c
This patch enables robust and complete testing of a compute scenario
where RBD is used for Cinder and Nova.
* Allow cinder.pp and nova.pp to configure RBD backend.
* Switch scenario001 to run Cinder and Nova with RBD backend.
* Update README with more documentation about Cinder & Nova backends.
* Configure volume_clear option to 'none' so volumes are quickly
deleted.
* Enable rbd_default_features to 15 to increase ceph performances.
* Increase compute build_interval in Tempest to 60s, useful for
Bootfromvolume test where instance deletion can take time with RBD
backend.
Depends-On: I69a7d40e7d1847be06a843986ace4f0602272fe1
Depends-On: I7302b89da5a995e779ec349ab0c0f519c69a3a98
Depends-On: Ic410cb66e7620b6ca6acbea38360d8dd890000c9
Change-Id: I5c8d9cf2ff8fc361553b3eed73b697ad87170434
* Configure Ceph to use 'cephx' authentification method.
* Manage Admin, Mon and Clients keys.
* Manage pools for Glance, Cinder and Nova, used for RBD backend.
Depends-On: Ied38afc03d68f7868b1d1320752a48989fa2d018
Change-Id: I23b1163faf3253daa74c57121cd8e5da4f04cff9
