Handle radosgw hosts placement with non-default cluster name

In cephadm-adopt.yml TASK "Update the placement of radosgw hosts"
does not handle when Ansible var cluster is something other than
"ceph", unless this patch is used.

Update module ceph_orch_apply to support optional cluster
parameter using the same style as in module ceph_config.
The command is only extended to inclue the new keyring
and config options if cluster name is not ceph.

This patch is necessary to migrate older clusters which were
deployed when custom names were supported.

Closes: https://issues.redhat.com/browse/RHCEPH-10442
Signed-off-by: John Fulton <fulton@redhat.com>
(cherry picked from commit 8b2f2133fab63950c92daa06851ef2d8680a3007)

Handle adoption when radosgw_address_block is comma delimited list

In cephadm-adopt.yml TASK "Update the placement of radosgw hosts"
passes module ceph_orch_apply embedded YAML via a block scalar.
This YAML creates a Ceph spec of service_type RGW. The networks
key of this spec supports either a list or a string. Without this
patch, the networks key of the spec will only contain a string.
With this patch a string is still set for the networks key, but
if Ansible var radosgw_address_block contains commas, then var
radosgw_address_block is split by those commas into a list and
the networks key of the spec will be set to a list.

Closes: https://issues.redhat.com/browse/RHCEPH-10418
Signed-off-by: John Fulton <fulton@redhat.com>
(cherry picked from commit b0707ae5d884666a7eedcbbdb29c7d57d514456a)

cephadm-adopt: fix ssl condition in rgw placement task

rgw daemons fail to start as they were looking for
ssl certificates while none were available.

Signed-off-by: Teoman ONAY <tonay@ibm.com>
(cherry picked from commit bf76bded4c3760a6610284cc477f642c8635f06d)

doc: CI readthedocs job fails

Deprecation of projects using Sphinx without an explicit configuration file

[1] https://about.readthedocs.com/blog/2024/12/deprecate-config-files-without-sphinx-or-mkdocs-config/

Signed-off-by: Teoman ONAY <tonay@ibm.com>

Don't try to set devices fact when osd_auto_discovery was skipped

Right now, under certain OS and Ansible versions, ie Rocky Linux and
ansible-core 2.17, `devices_check` variable is getting defined even if
task was skipped.

That results in set_fact to fail, as resulting variable has no `results`
key in it.

Structure of such variable looks like that:
```
"devices_check": {
    "changed": false,
    "false_condition": "osd_auto_discovery | default(False) | bool",
    "skip_reason": "Conditional result was False",
    "skipped": true
}
```

Checking for task not being skipped solves such issues.

Signed-off-by: Dmitriy Rabotyagov <noonedeadpunk@gmail.com>
(cherry picked from commit 9405558d03b633f8677d0f5f559028ca61ea4403)

pytest: python-version typo

Fixes python-version

Signed-off-by: Teoman ONAY <tonay@ibm.com>
(cherry picked from commit c9ffb2e70cc0cffe82344865c3aca4011c7ade77)

cephadm-adopt: Alertmanager placement count missing

Regression from #7576. Alertmanager placement count was missing
after migration to ceph_orch_apply module

Signed-off-by: Teoman ONAY <tonay@ibm.com>
(cherry picked from commit 31be49506118ece75371d0c86af794caee8bc051)

cephadm-adopt: custom prometheus port lost after adoption

If a custom Prometheus port was used before adoption, it was not
taken into account and default 9095 was set instead. Now custom
port is re-applied.

Fixes: https://bugzilla.redhat.com/show_bug.cgi?id=2242346
Signed-off-by: Teoman ONAY <tonay@ibm.com>
(cherry picked from commit b41b7bf869982a82d9ddddcef714a672c530b6f7)

cephadm-adopt: add cephadm ssh config option

Add task to set customized ssh config for cephadm

Signed-off-by: Amir Nikpour <amnik.free@gmail.com>
(cherry picked from commit e4ab45d756d3271af1429103a942c534e46fd4b9)

ceph_orch_apply: fix yaml error when multiple rgw deployed

ceph orch ls rgw --format=yaml returns multiple documents
when multiple rgw are installed, this was not handled
correctly.

Signed-off-by: Teoman ONAY <tonay@ibm.com>
(cherry picked from commit 823700bc1bc1f477cd58c61cded977a706a5c308)

radosgw_user: support caps on user

Supporting adding/removing caps on the user.

Signed-off-by: Seena Fallah <seenafallah@gmail.com>
(cherry picked from commit c1f1c1e900ea2558d6090bd5d40dc7e7d49f075e)

ceph-pool: add support updating crush rule of the pool

Support updating crush rule of the pool via crush rule name.

Signed-off-by: Seena Fallah <seenafallah@gmail.com>
(cherry picked from commit d6ac0cc120c5552dba8ceb04a9ef77fe027e6cb4)

devices: test devices before collecting on auto discovery

In some scenarios with NVMe, a device might be identified by
Ansible but could actually be a multipath device rather than an
actual device. We need to exclude these as Ceph cannot create
OSDs on them.

Signed-off-by: Seena Fallah <seenafallah@gmail.com>
(cherry picked from commit be9b45852459434184b576d85ad80f505ee7ea3f)

cephadm-adopt: fix "Update the placement of radosgw hosts" task

networks was at the wrong level in the spec file. Failed with
"got an unexpected keyword argument 'networks'"

Signed-off-by: Teoman ONAY <tonay@ibm.com>
(cherry picked from commit 2c88ecc7847d3c746fb7c324414d098470c54289)

cephadm-adopt: Fixes binding network for alertmanager

Alertmanager was bind to default * network instead of grafana_server_addr
as it was before. Now on if grafana_server_addr is defined, it will be
bind to that network.

Signed-off-by: Teoman ONAY <tonay@ibm.com>
(cherry picked from commit 0bf3398774dda6683f74a48f0ff3b26be6acec5a)

vagrant-box atomic-host -> stream9

Signed-off-by: Teoman ONAY <tonay@ibm.com>

cephadm-adopt: fix "Update the placement of radosgw hosts" task

spec file template conditions were incorrect

Signed-off-by: Teoman ONAY <tonay@ibm.com>
(cherry picked from commit e85060cb6713b05757f87f5bda3368416aa6e64e)

Replace Centos stream 8 by stream 9

In test scenarios

Signed-off-by: Teoman ONAY <tonay@ibm.com>

ceph-handler: use haproxy maintenance for rgw restarts

RGW currently restarts without waiting for existing connections to
close. By adjusting the HAProxy weight before the restart, we can
ensure that no active connections are disrupted during the restart
process.

Signed-off-by: Seena Fallah <seenafallah@gmail.com>
(cherry picked from commit 4fa9057a3ebb4c8929df23c1a2b12554eb3957e4)

Revert "nfs-ganesha support removal"

This reverts commit 675667e1d60b7080dad7293f2954de23718c5042.

Signed-off-by: Guillaume Abrioux <gabrioux@ibm.com>
(cherry picked from commit 59198f5bcdf2eca5cc99c25951d93513c508c01e)

ceph_orch_apply: fix idempotency

As is idempotency does not work as the ceph orch
output does contain more attributes than the expected
spec.

Signed-off-by: Teoman ONAY <tonay@ibm.com>
(cherry picked from commit 5a44fbeaa69273260f42257524df2e211b65c3ce)

ceph_orch_spec: Add ceph orch apply spec feature

Add new module ceph_orch_spec which applies ceph spec files.
This feature was needed to bind extra mount points to the RGW
container (/etc/pki/ca-trust/).

Fixes: https://bugzilla.redhat.com/show_bug.cgi?id=2262133
Signed-off-by: Teoman ONAY <tonay@ibm.com>
(cherry picked from commit f6fd034e7ee167ea31a382c30432972aeb7adba5)

ceph-key: check for key existance on absent state

check if the key is exists before removal.

Signed-off-by: Seena Fallah <seenafallah@gmail.com>
(cherry picked from commit 52365166a4b3757d174b602891fb54bc751defe0)

ceph-container: use ceph user and group

Use --setuser and --setgroup to ceph to run daemons with ceph user.

Signed-off-by: Seena Fallah <seenafallah@gmail.com>
(cherry picked from commit 7016c6df3cf5fc28f8b931ef9083de9dcf5e6e9c)

tests: collect mons and mgrs logs

Signed-off-by: Seena Fallah <seenafallah@gmail.com>
(cherry picked from commit 779a13eb31db10cff0ca315b9907ed68a6f480a1)

ceph-container: keep run dir permission consistent

Make it the same as "Create ceph initial directories" task
and make owner and group 167 for containers so they can write
with ceph user.

Signed-off-by: Seena Fallah <seenafallah@gmail.com>
(cherry picked from commit 65c84a3583ea4cceb3b4baf70757e62c70a5ccea)

ceph-rgw: introduce rgw zone to the name schema

This is needed by ceph-exporter as it is parsing the socket by the number of dots.
Although the rgw_zone variable is only using for constructing the client name
and has nothing to do with multisiting.

Signed-off-by: Seena Fallah <seenafallah@gmail.com>
(cherry picked from commit 1121e6d98a41a913886240bcb873380008afcafc)

Fix cephadm-adopt test scenario

Fixes cephadm-adopt test scenario by configuring rbd application
on test and test2 pools. Otherwise cephadm-adopt failed at task
"Check pools have an application enabled"

Signed-off-by: Teoman ONAY <tonay@ibm.com>
(cherry picked from commit ca191fad2ab45d4c4217251078efcfe965898d6a)

radosgw_user: handle non-existence user on check mode

Set return code to 0 on check mode when the user doesn't exists and it will fail the module.

Signed-off-by: Seena Fallah <seenafallah@gmail.com>
(cherry picked from commit 878cce5b4847a9a112f9d07c0fd651aa15f1e58b)

ceph-mon: move the ceph_config call to ceph-mon role

ba7eb62a1bcbd925b053cbd737012f034568a6f7 broke deployments where
mgrs are not collocated with mons.

Signed-off-by: Guillaume Abrioux <gabrioux@ibm.com>
(cherry picked from commit 4f6641fe5ca31346ef6c2710b699f7b9e5319982)

ceph-config: move `ceph_config` module call to ceph-config role

so you don't have to implement a similar play on your own
if not using main (sample) playbooks.

Signed-off-by: Guillaume Abrioux <gabrioux@ibm.com>
(cherry picked from commit ba7eb62a1bcbd925b053cbd737012f034568a6f7)

ceph_volume: set changed to false to readonly actions

Signed-off-by: Seena Fallah <seenafallah@gmail.com>
(cherry picked from commit fb13ee35bf2cf096db3f66b516eaa7ffa0d24770)

ceph_crush: add support check mode

Signed-off-by: Seena Fallah <seenafallah@gmail.com>
(cherry picked from commit 326e2e4d6b66d5a27dfa64cd6d0ea892c0f3b119)

ceph-mon: fix admin keyring creation task

Although custom cluster name support was dropped, it breaks ceph-volume
functional testing as it uses "test" as cluster name.
The plan is to make ceph-volume use "ceph" but for now it's easier to
address the issue in this task.

Signed-off-by: Guillaume Abrioux <gabrioux@ibm.com>
(cherry picked from commit c4b3520730612da10e6d7f3c61ce0a5446725e47)

common: install python3-packaging on centos el8

ceph-volume has a dependency on `python3-packaging` which is available
in PowerTools repo.

Signed-off-by: Guillaume Abrioux <gabrioux@ibm.com>
(cherry picked from commit 8ebff2124db1e670099b7e6d141d6c3178c3c181)

ceph-handler: allow 405 as a success restart for rgw

If rgw is configured to only operate admin api - this is the status code it will return for GET on no path.

Signed-off-by: Seena Fallah <seenafallah@gmail.com>
(cherry picked from commit 9f9b32e9e6610e62aabeda4c09339ac3e84e7c4d)

ceph_pool: add support check mode

Signed-off-by: Seena Fallah <seenafallah@gmail.com>
(cherry picked from commit a45defd3869e95b06730abc0fe741c290d632b8c)

handlers: move tmpdir removal

This moves the tmpdir removal tasks to main playbook.
The handlers in Ansible are too restrictive and doesn't work
quite well for our use case.

Signed-off-by: Guillaume Abrioux <gabrioux@ibm.com>
(cherry picked from commit bab034024c7dfdbce25531963c162dd94e4f394d)

nfs-ganesha support removal

nfs-ganesha support will be implemented in a separate playbook.

Signed-off-by: Guillaume Abrioux <gabrioux@ibm.com>
(cherry picked from commit 675667e1d60b7080dad7293f2954de23718c5042)

vagrantfile: drop grafana group

this has never been used for real, let's drop this

Signed-off-by: Guillaume Abrioux <gabrioux@ibm.com>
(cherry picked from commit 1ebddeccb38a34d66f96f0c34db2b6958685e5e1)

remove legacy task

this was for backward compatibility concerns, it's time to drop this
task.

Signed-off-by: Guillaume Abrioux <gabrioux@ibm.com>
(cherry picked from commit d9e2de1d1d84fee877707460328a232512680025)

common: configure epel repository (rockylinux)

this is needed in order to make some libraries available when
running on RockyLinux

Signed-off-by: Guillaume Abrioux <gabrioux@ibm.com>
(cherry picked from commit 867be7c35c1f10080f842c3c6bdd51062cfd788a)

tests: various changes

some minor changes

Signed-off-by: Guillaume Abrioux <gabrioux@ibm.com>
(cherry picked from commit c42c0c355884cec984e21d49119c5b84fc489340)

rgw_profiles: update rgw-keystone profiles

See issue #7458

Fixes: #7458
Signed-off-by: Guillaume Abrioux <gabrioux@ibm.com>
(cherry picked from commit 0f341cc68bdce7547ff06c477d926c82e3ff2880)

container-engine: enforce docker for Ubuntu os family

This enforces docker.io and docker respectively for
`container_package_name` and `container_service_name` by default
for Ubuntu distribution.

Fixes: #7496
Signed-off-by: Guillaume Abrioux <gabrioux@ibm.com>
(cherry picked from commit ef5a09d0e041806369a6caa7f7d062a47811a08f)

ceph_config: compare current vals in low

Check the current value and asked both in low case as the asked one might be in capital as it should be.

Signed-off-by: Seena Fallah <seenafallah@gmail.com>
(cherry picked from commit ed4b6e71b3b832848e326ccafd9ba4fb036a5f8c)

ceph_config: add support config remove

Add support rm action.

Signed-off-by: Seena Fallah <seenafallah@gmail.com>
(cherry picked from commit fa2eb2532e1f4b8d4a9ea9280ffa9ad3d1cf2a0e)

radosgw_user: add support check mode

Signed-off-by: Seena Fallah <seenafallah@gmail.com>
(cherry picked from commit a7e881439ec9aa70fb6e8964d414ff77e092ebb3)

radosgw_user: parse system and admin as boolean

The returned payload from rgw has them as a boolean. By having them as a string it would always report a change and try to modify the user.

Signed-off-by: Seena Fallah <seenafallah@gmail.com>
(cherry picked from commit a95726c40900d30bb3728fc99e7487c65ea2d0a1)

ceph_ec_profile: make int params str

Ceph return them as string so in comparision it would always fail

Signed-off-by: Seena Fallah <seenafallah@gmail.com>
(cherry picked from commit f0b4cc3f0d6c82ec6433935bfd443367422c76a5)

ceph-osd: introduce ec profile creation

RGW Pools can now use the existing ec profiles and rules created by ceph-osd role.

Signed-off-by: Seena Fallah <seenafallah@gmail.com>
(cherry picked from commit bbc1ba5c05d9a4444617e1777b10e127b345afb3)

ceph_ec_profile: add support more plugins

Add support more plugins with their respective args.

Signed-off-by: Seena Fallah <seenafallah@gmail.com>
(cherry picked from commit 80666a20dad82c6c0f8ad6a99c26bf52781bcfe3)

radosgw_zonegroup: add support check mode

Signed-off-by: Seena Fallah <seenafallah@gmail.com>
(cherry picked from commit beb046cba9498bc3febb282efef43a32c09c200b)

radosgw_zonegroup: parse master as boolean

The returned payload from rgw has is_master as a boolean. By having master as a string it would always report a change and try to modify the zonegroup.

Signed-off-by: Seena Fallah <seenafallah@gmail.com>
(cherry picked from commit e8d34b6ee6c51a68be1ee6ed63485699a4d07b9a)

ceph-config: check for set_radosgw_address before importing task

Signed-off-by: Seena Fallah <seenafallah@gmail.com>
(cherry picked from commit 49ec2479ea347d29616feec52386562dfa406676)

ceph-handler: accept 404 as a success status for rgw restart

404 is the status code returned by rgw when it is serving website api. This needs to be consider as a healthy status for rgw.

Signed-off-by: Seena Fallah <seenafallah@gmail.com>
(cherry picked from commit f96c0441e3cd33c3f5ff66dfac80d69b5d609dd8)

radosgw_zone: add support zone set

Support zone set from a json doc

Signed-off-by: Seena Fallah <seenafallah@gmail.com>
(cherry picked from commit e219892aebf07e3aa941a69461c228e91f7066a2)

ceph-volume: add support check mode

Run read-only actions on check mode

Signed-off-by: Seena Fallah <seenafallah@gmail.com>
(cherry picked from commit b60e0134a5d461fcbcfa5471992487ad88876d2e)

Revert "add a warning regarding current status of stable-8.0"

This reverts commit 4e1e61da15375b2343df6b66e0916f319f1c52dc.

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

tests/upgrade: various changes

- use el8 image,
- disable nfs-ganesha testing

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

tests: run vagrant_up.sh from dirname

Signed-off-by: Guillaume Abrioux <gabrioux@ibm.com>
(cherry picked from commit 3c0e06ea0cdf3098497274dd8ff5a63b117d20c5)

tests: enable nfs-ganesha testing

This makes the container-upgrade scenario fail as this gets deployed
during initial quincy deployment but doesn't get upgraded to reef

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

tests: skip ceph_crash and ceph_exporter tests post upgrade

as stable-8.0 is dropping iscsigws support, after the upgrade to reef
from quincy these tests should be skipped.

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

update: set cluster config prior to upgrading

As 8.0 drops cluster config from ceph.conf by default, we need
to make sure this config is well applied as part of the upgrade to 8
(reef) from 7 (quincy)

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

tests: update --ssh-config parameter (tox-update.ini)

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

update: ignore linter errors on ceph-exporter systemd tasks

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

replace main references with reef

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

validate: check for reef repository

stable-8.0 is reef so this must check the right value for ceph_stable_release

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

osd: drop openstack related tasks

All of this should be addressed in custom separate playbooks if needed.

Signed-off-by: Guillaume Abrioux <gabrioux@ibm.com>
(cherry picked from commit 9c467e41b39af63dec9a0b237b550edb9c57da85)

tests: drop ooo-collocation legacy files

this job was removed, these files are no longer needed, let's remove
them.

Signed-off-by: Guillaume Abrioux <gabrioux@ibm.com>
(cherry picked from commit 05a1465416593d922e600192e4be80823ff0f0d7)

ceph-mon: add missing no_log for admin keyring tasks

Signed-off-by: Seena Fallah <seenafallah@gmail.com>

group_vars: resync samples with default values

2b72ea991d40fc12f277edd58825ce98f068e8c4 and
c8eeae243efac9ad894f3240a8a256b05336f75b have missed the resync.

This addresses that.

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

simplify monitor address setting

this drops the following parameters:

- monitor_address_block
- monitor_interface
- monitor_address

The monitor address will be automatically set from `public_network` parameter.

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

fix upgrade

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

rolling_update: address linter errors

this addresses 3 linter errors in rolling update playbook

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

add a warning regarding current status of stable-8.0

stable-8.0 is not ready and shouldn't be used for now.

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

tests: add ceph exporter to collect logs

Signed-off-by: Seena Fallah <seenafallah@gmail.com>

ceph-exporter: add installation role

Signed-off-by: Seena Fallah <seenafallah@gmail.com>

mons: use hostname for initial members

Signed-off-by: Seena Fallah <seenafallah@gmail.com>

systemd: export params as a varaible

This can help to have extra params or modify the existing ones via group vars.

Signed-off-by: Seena Fallah <seenafallah@gmail.com>

site: install mgrs with mons if sharing the same host

If mgr is meant to be installed on the mon host it needs to be installed in the same playbook as restart handlers might failed because of non-existance mgr

Signed-off-by: Seena Fallah <seenafallah@gmail.com>

ceph-volume: disable dmcrypt by default

regression by dcdb1710cbaf7946cf8161b4395697d4898680d3

Signed-off-by: Seena Fallah <seenafallah@gmail.com>

ceph-facts: make set_radosgw_address optional

This can help to define custom rgw_instances with custom names and ports and addresses.

Signed-off-by: Seena Fallah <seenafallah@gmail.com>

Welcome Back :tada:

Signed-off-by: Seena Fallah <seenafallah@gmail.com>

tox: enable diff for ansible runs

Diff option can help in CI when a task fail because of a corrupted render in previous tasks.

Signed-off-by: Seena Fallah <seenafallah@gmail.com>

container: cleanup container systemd units

* Make common params of container args in a var to avoid duplication
* The /var/lib/ceph/crash mount was missing after 637ca81c9cf801e4d1d125dc8a2492b90fd78eea
* Add CEPH_USE_RANDOM_NONCE as it's needed when running inside container (can be removed for squid later)
* Add NODE_NAME as some part of ceph code relies on this var
* add default logging opts for

Signed-off-by: Seena Fallah <seenafallah@gmail.com>

refactor: remove multisite leftovers

Signed-off-by: Seena Fallah <seenafallah@gmail.com>

ceph-handler: restart existed rgw daemons

This is needed for new instances are the restart might trigger before the deployment

Signed-off-by: Seena Fallah <seenafallah@gmail.com>

ceph-handler: remove tempdir when all handlers are done

Signed-off-by: Seena Fallah <seenafallah@gmail.com>

ceph-config: make rgw config to be in file

Signed-off-by: Seena Fallah <seenafallah@gmail.com>

ceph-config: introduce dedicated cluster config flow

Signed-off-by: Seena Fallah <seenafallah@gmail.com>

drop rhcs references

RHCS moved away from ceph-ansible. All RHCS references should be
removed.

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

drop iscsigw support

This service is no longer maintained.
Let's drop its support within ceph-ansible.

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

address Ansible linter errors

This addresses all errors reported by the Ansible linter.

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

drop rgw multisite deployment support

The current approach is extremely complex and introduced a lot
of spaghetti code. This doesn't offer a good user experience at all.

It's time to think to another approach (dedicated playbook) and drop
the current implementation in order to clean up the code.

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

tests: set ANSIBLE_COLLECTIONS_PATH for rbdmirror scenario

This is missing so the job fails.
Typical error:
```
ERROR! couldn't resolve module/action 'community.general.dnf_config_manager'.
```

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

facts: remove legacy tasks

these tasks were there only for backward compatibility concerns.
It's time to drop them.

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

bump ansible-core to 2.16

- Remove python3 shebangs
- command module warn parameter is deprecated since 2.11 and removed
  from 2.14

Signed-off-by: Teoman ONAY <tonay@ibm.com>

tests: bump to py3.10

let's bump the python version to 3.10 in python-based
github actions.

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

pytest-infra: fix failures

pytest-infra 10.0.0 release changed the get_variables() behavior which led to CI failures
(https://github.com/pytest-dev/pytest-testinfra/pull/724)
Starting of now all groups the host is part of and not only the direct parent groups are
returned by get_variables()

Signed-off-by: Teoman ONAY <tonay@ibm.com>