When using community repository we need to set the priority on the
ceph repositories because we could have some conflict with EPEL
packages.
In order to set the priority on the ceph repositories, we need to
install the yum-plugin-priorities package.

http://docs.ceph.com/docs/master/install/get-packages/#rpm-packages



Signed-off-by: Dimitri Savineau <dsavinea@redhat.com>

Instead of using subscription-manager with command module we can use
the rhsm_repository ansible module.
This module already uses repos list feature to determine if a
repository is enabled or not. That way this module is idempotent so
we don't need changed_when: false anymore.

Signed-off-by: Dimitri Savineau <dsavinea@redhat.com>

The ceph stable community repository only enables the basearch
packages url.
Adding the noarch url because starting with nautilus release, some
packages are added there and useful for mgr or grafana.

Signed-off-by: Dimitri Savineau <dsavinea@redhat.com>

We shouldn't reset `ceph_release` with `ceph_stable_release` when
`ceph_repository` is `rhcs`

Closes: https://bugzilla.redhat.com/show_bug.cgi?id=1645379



Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>

When {{omit}} is concatenated with another string, it expands to something
like __omit_place_holder__63eea0d96dd6ed867b95405e11d87dddf61f448d.
However, in these use-cases we need an empty string.

Regression introduced in d53f55e8.

Signed-off-by: Leah Neukirchen <leah.neukirchen@mayflower.de>

let's not be blocked by the fact we don't have the initial keyring in
`{{ fetch_directory }}`

Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>

When we're deploying a dev branch.

Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>

when `ceph_origin` is set to `'repository'` and `ceph_repository` to
`'community'` we need to ensure `ceph_release` reflect
`ceph_stable_release`.

4a3f180f9d29d5a31468ebb3d1c5f31a53a93960 simply removed the override
while it should just have to be run only when the condition mentioned
above is satisfied.

Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>

This condition is useless and it's also creating issues we don't see in
our CI. ceph_release is set by either ceph-common or ceph-docker-common
so let's keep it this way.

Closes: https://bugzilla.redhat.com/show_bug.cgi?id=1645379



Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>

Make linter happy and add more robustness to remote tasks by retrying 3
times (the default) before failing.

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

add register/until on all packaging related tasks to avoid non valid CI
failure.

Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>

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

Use delegate_to: localhost instead.

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

Fix tasks using variables that did not have space before and after {{
  }}

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

The use of a handler meant that the cache would be updated at the very
end of the play, which doesn't work when adding a development repo and
trying to install right after it. This mostly reverts
53cdddf88699263763b36643565e5f846d9d13a8 without an actual `git revert`
because that caused other conflicts.

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

The task was included by mistake while resolving a merge confict for
commit 8edbda96df6896d51703fcc250f562abb4011a2d. The task removed in
commit b3a71eeb08e9cdb2607ed60d724f387a0a24d3de.

Fixes: https://github.com/ceph/ceph-ansible/issues/3292


Signed-off-by: Rishabh Dave <ridave@redhat.com>

Update the meta with the relavant support such as:

* ansible version: min 2.4
* distro supported (tested on) centos 7

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

Fixes : [error] too many blank lines (1 > 0)

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

Use yum module to list repos and then activate them if needed.

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

We need a handler because the task changed, the old implementation was
basically mimicing a handler.

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

Calling command should have changed_when false otherwise each time it
runs it will show as 'changed' and this is irrelevant.
Commands should not change things if nothing needs doing

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

Using block directives simplifies the playbooks and makes them more
readable.

Fixes: https://github.com/ceph/ceph-ansible/issues/2835


Signed-off-by: Rishabh Dave <ridave@redhat.com>

Fix [error] too many blank lines (1 > 0) (empty-lines)

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

Move the handler to restart firewall in ceph-infra role.

Closes: #3243

Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>

this role manages ceph infra services such as ntp, firewall, ...

Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>

Fixes the deprecation warning:

  [DEPRECATION WARNING]: Using tests as filters is deprecated. Instead of
  using `result|search` use `result is search`.

Signed-off-by: Noah Watkins <nwatkins@redhat.com>

Check firewall isn't working as expected and might break deployments.
This part of the code will be reworked soon.

Let's focus on configure_firewall code for now.

Closes: https://bugzilla.redhat.com/show_bug.cgi?id=1541840



Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>

Instead used "import_tasks" and "include_tasks" to tell whether tasks
must be included statically or dynamically.

Fixes: https://github.com/ceph/ceph-ansible/issues/2998


Signed-off-by: Rishabh Dave <ridave@redhat.com>

Allow user to choose between timesyncd, chronyd and ntpd
Installation will default to timesyncd since it is distributed as
part of the systemd installation for most distros.
Added note indicating NTP daemon type is not used for containerized
deployments.

Fixes issue #3086 on Github

Signed-off-by: Benjamin Cherian <benjamin_cherian@amat.com>

Use "import_tasks" or "include_tasks" instead.

Signed-off-by: Rishabh Dave <ridave@redhat.com>

A couple if things were wrong in the initial commit:

* ceph_release_num[ceph_release] >= ceph_release_num['luminous'] will
never work since the ceph_release fact is set in the roles after. So
either ceph-common or ceph-docker-common set it

* we can easily re-use the initial command to check if a cluster is
running, it's more elegant than running it twice.

* set the fact rgw_hostname on rgw nodes only

Closes: https://bugzilla.redhat.com/show_bug.cgi?id=1618678


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

The config_template plugin exists in the ceph-common role so that
config_template will still work with ansible galaxy.

This PR syncs the config_template module from the base of the repo in
plugins/actions to the ceph-common role.

Signed-off-by: Andy McCrae <andy.mccrae@gmail.com>

Since the container now simply reads the ceph.conf, we remove all the
unnecessary options.

Also this PR is the foundation to support multiple backend, such as the
new 'beast' from Ceph Mimic.

Closes: https://bugzilla.redhat.com/show_bug.cgi?id=1582411


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

We forgot to add mgr_group_name when checking for the mon repo, thus the
conditional on the next task was failing.

Closes: https://bugzilla.redhat.com/show_bug.cgi?id=1598185


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

Signed-off-by: Vasu Kulkarni <vasu@redhat.com>

During 226f80c22bf61a7e8f00f4cca5f35eda67280250 only Debian package
installs had the correct state set to ensure packages were upgraded when
the "upgrade_ceph_packages" var was set to true.

Signed-off-by: Andy McCrae <andy.mccrae@gmail.com>

Currently we expect that if configure_firewall is set to True to have
firewalld enabled and running. Let's enforce that.

Closes: https://bugzilla.redhat.com/show_bug.cgi?id=1589146


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

Prior to this patch if you were running on a Red Hat system,
ceph-ansible would try to configure firewalld for you without the
operators's consent.
Now you can enable or disable the fw configuration by setting
configure_firewall to either true or false.

Closes: https://bugzilla.redhat.com/show_bug.cgi?id=1589146


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

We need to do this because on dev or rhcs installs ceph_stable_release
is not mandatory and the firewall check tasks have a task that is
conditional based off the installed version of ceph. If we perform those
checks after package install then they will not fail on dev or rhcs
installs.

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

Prior to this commit the firewall tasks were not opening the ceph-mgr
ports. This would lead to unclean configuration since the ceph-mgr
daemons can not connect to the OSDs.
Thi commit opens the right ports on the ceph-mgr nodes to talk with the
OSDs.

Closes: https://bugzilla.redhat.com/show_bug.cgi?id=1526400


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