]> git.apps.os.sepia.ceph.com Git - ceph-ansible.git/log
ceph-ansible.git
8 years agodocker: we don't need to copy the ceph.conf on all the nodes 1929/head
Sébastien Han [Wed, 20 Sep 2017 14:33:29 +0000 (16:33 +0200)]
docker: we don't need to copy the ceph.conf on all the nodes

We generate the ceph.conf on all the nodes through the
ceph-docker-common so there is no need to push it to the Ansible file.

Also this is breaking the ceph.conf template generation since we only
generate sections based on the host the ansible task is running on.

For example, what's typically happening, we bootstrap the monitor, we
get a ceph.conf generated for a mon only, we go on an osd, we generate
the ceph.conf with osd section (done by ceph-docker-common) but this
gets overwritten by the copy_config task of the ceph-osd role.

Signed-off-by: Sébastien Han <seb@redhat.com>
8 years agoMerge pull request #1920 from jprovaznik/ganesha
Sébastien Han [Wed, 20 Sep 2017 12:48:36 +0000 (14:48 +0200)]
Merge pull request #1920 from jprovaznik/ganesha

Make ceph-nfs service enablement/start optional

8 years agoMerge pull request #1926 from ceph/rhcs-debina v3.0.0rc10
Sébastien Han [Tue, 19 Sep 2017 17:50:40 +0000 (19:50 +0200)]
Merge pull request #1926 from ceph/rhcs-debina

common: fix rhcs debian install

8 years agocommon: fix rhcs debian install 1926/head
Sébastien Han [Tue, 19 Sep 2017 17:49:44 +0000 (19:49 +0200)]
common: fix rhcs debian install

Closes: https://bugzilla.redhat.com/show_bug.cgi?id=1493231
Signed-off-by: Sébastien Han <seb@redhat.com>
8 years agoMerge pull request #1925 from ceph/wip-nfs-config-changes
Sébastien Han [Tue, 19 Sep 2017 16:57:24 +0000 (18:57 +0200)]
Merge pull request #1925 from ceph/wip-nfs-config-changes

nfs: ganesha.conf template fixes

8 years agonfs: ganesha.conf template fixes 1925/head
Ali Maredia [Tue, 19 Sep 2017 16:45:24 +0000 (12:45 -0400)]
nfs: ganesha.conf template fixes

- Change capitalization of config options to be
in line with what config.txt in the nfs-ganesha
tree says

Signed-off-by: Ali Maredia <amaredia@redhat.com>
8 years agoMerge pull request #1918 from ceph/fix-rhcinstall
Sébastien Han [Tue, 19 Sep 2017 10:58:52 +0000 (12:58 +0200)]
Merge pull request #1918 from ceph/fix-rhcinstall

common: fix rhcs installation and rgw package for nfs

8 years agocommon: fix rhcs installation and rgw package for nfs 1918/head
Sébastien Han [Mon, 18 Sep 2017 21:56:38 +0000 (23:56 +0200)]
common: fix rhcs installation and rgw package for nfs

RHCS install wasn't working at all prior to this commit as the name of
the include was pointing to a non-existing file.

Closes: https://bugzilla.redhat.com/show_bug.cgi?id=1492056
Signed-off-by: Sébastien Han <seb@redhat.com>
8 years agoMerge pull request #1919 from ceph/iscsi-check
Sébastien Han [Tue, 19 Sep 2017 10:10:10 +0000 (12:10 +0200)]
Merge pull request #1919 from ceph/iscsi-check

common: fix rhel check

8 years agoMerge pull request #1904 from ceph/name-include-fact
Sébastien Han [Tue, 19 Sep 2017 10:09:25 +0000 (12:09 +0200)]
Merge pull request #1904 from ceph/name-include-fact

name includes and set_fact for clarity

8 years agoMake ceph-nfs service enablement/start optional 1920/head
Jan Provaznik [Tue, 19 Sep 2017 09:59:54 +0000 (11:59 +0200)]
Make ceph-nfs service enablement/start optional

When ceph-nfs service is managed by pacemaker, it's useful to
not enable and start ceph-nfs service through systemd but let
pacemaker to start the service in a next step.

8 years agocommon: fix rhel check 1919/head
Sébastien Han [Tue, 19 Sep 2017 09:55:26 +0000 (11:55 +0200)]
common: fix rhel check

Looks like Ansible is now using "RedHat" instead of "Red Hat Enterprise
Linux"

Signed-off-by: Sébastien Han <seb@redhat.com>
8 years agoMerge pull request #1911 from fghaas/1910
Sébastien Han [Tue, 19 Sep 2017 08:03:46 +0000 (10:03 +0200)]
Merge pull request #1911 from fghaas/1910

Introduce ceph_nfs_ceph_user

8 years agoIntroduce ceph_nfs_ceph_user 1911/head
Florian Haas [Fri, 15 Sep 2017 15:38:19 +0000 (17:38 +0200)]
Introduce ceph_nfs_ceph_user

In analogy to ceph_nfs_rgw_user, we should be able to define a user
with which the nfs-ganesha Ceph FSAL connects to the cluster.

Introduce a ceph_nfs_ceph_user variable, setting its default to
"admin" (which preserves the prior behavior of always connecting as
client.admin).

Fixes #1910.

8 years agoname includes and set_fact for clarity 1904/head
Sébastien Han [Thu, 14 Sep 2017 22:48:53 +0000 (16:48 -0600)]
name includes and set_fact for clarity

When Ansible is not run with verbose options it's difficult to see which
include and/or set_fact does what. So adding a name for each clarifies.

Signed-off-by: Sébastien Han <seb@redhat.com>
8 years agoMerge pull request #1889 from ceph/client-containers v3.0.0rc9
Sébastien Han [Mon, 18 Sep 2017 15:27:32 +0000 (17:27 +0200)]
Merge pull request #1889 from ceph/client-containers

client: ability to create keys and pool with no ceph binaries

8 years agoMerge pull request #1915 from ceph/state-leftover
Sébastien Han [Mon, 18 Sep 2017 13:46:07 +0000 (15:46 +0200)]
Merge pull request #1915 from ceph/state-leftover

docker-common: re-introduce state for leftover files

8 years agodocker-common: re-introduce state for leftover files 1915/head
Sébastien Han [Mon, 18 Sep 2017 13:01:32 +0000 (15:01 +0200)]
docker-common: re-introduce state for leftover files

The variable "statleftover" was removed by commit
a60c74f61eb4b7523ef6d4065624312bfd0aa03a
and never added back to the new playbook,
yet it is still being referenced.

Adding it back

Closes: https://bugzilla.redhat.com/show_bug.cgi?id=1492224
Signed-off-by: Sébastien Han <seb@redhat.com>
8 years agoMerge pull request #1900 from ceph/rpm-all-yaml
Sébastien Han [Mon, 18 Sep 2017 12:52:11 +0000 (14:52 +0200)]
Merge pull request #1900 from ceph/rpm-all-yaml

rpm: ship rhcs.yaml.sample as all.yml.sample on RHEL

8 years agoclient: ability to create keys and pool with no cpeh binaries 1889/head
Sébastien Han [Wed, 13 Sep 2017 20:13:53 +0000 (14:13 -0600)]
client: ability to create keys and pool with no cpeh binaries

On a container env, machines don't have any ceph binaries so we need to
use a container to run the commands.

Signed-off-by: Sébastien Han <seb@redhat.com>
8 years agoMerge pull request #1891 from ceph/clarify-update
Sébastien Han [Fri, 15 Sep 2017 13:08:49 +0000 (07:08 -0600)]
Merge pull request #1891 from ceph/clarify-update

rolling_update: clarify update doc

8 years agoMerge pull request #1898 from ceph/restart-mon
Sébastien Han [Fri, 15 Sep 2017 12:23:51 +0000 (06:23 -0600)]
Merge pull request #1898 from ceph/restart-mon

defaults: restart docker daemon higher delay

8 years agoMerge pull request #1901 from ceph/rpm-ansible-2.3
Sébastien Han [Thu, 14 Sep 2017 23:05:08 +0000 (17:05 -0600)]
Merge pull request #1901 from ceph/rpm-ansible-2.3

rpm: require Ansible 2.3.1.0

8 years agorpm: require Ansible 2.3.1.0 1901/head
Ken Dreyer [Thu, 14 Sep 2017 22:12:06 +0000 (16:12 -0600)]
rpm: require Ansible 2.3.1.0

We require a newer verison of Ansible now.

8 years agorpm: ship rhcs.yaml.sample as all.yml.sample on RHEL 1900/head
Ken Dreyer [Thu, 14 Sep 2017 20:42:53 +0000 (14:42 -0600)]
rpm: ship rhcs.yaml.sample as all.yml.sample on RHEL

When building the RPM on RHEL, we will ship the conents of the
`rhcs.yml.sample` file into `all.yml.sample`.

The purpose of this change is to expose "all.yml.sample" to both
upstream and downstream users, so they use the same file name (but with
different contents.)

8 years agoMerge pull request #1899 from ceph/rpm-obsoletes-iscsi
Sébastien Han [Thu, 14 Sep 2017 20:18:02 +0000 (14:18 -0600)]
Merge pull request #1899 from ceph/rpm-obsoletes-iscsi

rpm: Obsoletes ceph-iscsi-ansible

8 years agorpm: Obsoletes ceph-iscsi-ansible 1899/head
Ken Dreyer [Thu, 14 Sep 2017 20:12:09 +0000 (14:12 -0600)]
rpm: Obsoletes ceph-iscsi-ansible

The ceph-iscsi-ansible project has moved into ceph-ansible now. Make the
ceph-ansible package obsolete the ceph-iscsi-ansible package. With this
change, Yum updates will seamlessly uninstall the ceph-iscsi-ansible
package.

8 years agoMerge pull request #1895 from ceph/rhcs-on-rhel
Sébastien Han [Thu, 14 Sep 2017 19:44:05 +0000 (13:44 -0600)]
Merge pull request #1895 from ceph/rhcs-on-rhel

rpm: ship different yml samples on RHEL vs CentOS

8 years agodefaults: restart docker daemon higher delay 1898/head
Sébastien Han [Thu, 14 Sep 2017 19:38:11 +0000 (13:38 -0600)]
defaults: restart docker daemon higher delay

Use default delay since the mon (in particular) can take more time to
restart.
Solves error with:

STDERR:

Error response from daemon: No such container: ceph-mon-mon0

Signed-off-by: Sébastien Han <seb@redhat.com>
8 years agoMerge pull request #1894 from ceph/mgr-version
Andrew Schoen [Thu, 14 Sep 2017 18:47:09 +0000 (13:47 -0500)]
Merge pull request #1894 from ceph/mgr-version

mon: create the mgr key for release >= luminous

8 years agorpm: ship different yml samples on RHEL vs CentOS 1895/head
Ken Dreyer [Thu, 14 Sep 2017 17:22:37 +0000 (11:22 -0600)]
rpm: ship different yml samples on RHEL vs CentOS

When building the RPM on CentOS, we will ship the all.yml.sample file
and delete rhcs.yml.sample.

When building the RPM on RHEL, we will ship the rhcs.yml.sample file and
delete all.yml.sample.

8 years agomon: create the mgr key for release >= luminous 1894/head
Sébastien Han [Thu, 14 Sep 2017 17:06:44 +0000 (11:06 -0600)]
mon: create the mgr key for release >= luminous

This fixes RHCS builds. We know which Ceph version we are running on.

Signed-off-by: Sébastien Han <seb@redhat.com>
8 years agoMerge pull request #1893 from ceph/rgw-socker
Sébastien Han [Thu, 14 Sep 2017 16:52:30 +0000 (10:52 -0600)]
Merge pull request #1893 from ceph/rgw-socker

rgw: remove old variables

8 years agoMerge pull request #1890 from ceph/lvm-setup
Andrew Schoen [Thu, 14 Sep 2017 16:38:13 +0000 (11:38 -0500)]
Merge pull request #1890 from ceph/lvm-setup

tests: fix lvm_setup.yml for purge_cluster.yml

8 years agorgw: remove old variables 1893/head
Sébastien Han [Thu, 14 Sep 2017 15:42:50 +0000 (09:42 -0600)]
rgw: remove old variables

Since the only support civetweb these variables are obsolete.

Signed-off-by: Sébastien Han <seb@redhat.com>
8 years agoosd: add meaningful message for journal_size
Sébastien Han [Thu, 14 Sep 2017 05:48:41 +0000 (23:48 -0600)]
osd: add meaningful message for journal_size

Signed-off-by: Sébastien Han <seb@redhat.com>
8 years agoMerge pull request #1892 from ceph/purge-dmcrypt-col v3.0.0rc8
Sébastien Han [Wed, 13 Sep 2017 23:57:06 +0000 (17:57 -0600)]
Merge pull request #1892 from ceph/purge-dmcrypt-col

purge: only purge specific directories for mon

8 years agoMerge pull request #1800 from ceph/wip-osd-start-fix
Sébastien Han [Wed, 13 Sep 2017 23:20:10 +0000 (17:20 -0600)]
Merge pull request #1800 from ceph/wip-osd-start-fix

ceph-osd: Fix osd start sequence

8 years agopurge: only purge specific directories for mon 1892/head
Sébastien Han [Wed, 13 Sep 2017 23:07:04 +0000 (17:07 -0600)]
purge: only purge specific directories for mon

Handles the case when a mon is collocated with an OSD.

Closes: https://github.com/ceph/ceph-ansible/issues/1877
Signed-off-by: Sébastien Han <seb@redhat.com>
8 years agoMerge pull request #1885 from ceph/shrink-osd
Sébastien Han [Wed, 13 Sep 2017 22:12:49 +0000 (16:12 -0600)]
Merge pull request #1885 from ceph/shrink-osd

shrink-osd: fix when multiple osds

8 years agorolling_update: clarify update doc 1891/head
Sébastien Han [Wed, 13 Sep 2017 21:46:29 +0000 (15:46 -0600)]
rolling_update: clarify update doc

Closes: https://bugzilla.redhat.com/show_bug.cgi?id=1490188
Signed-off-by: Sébastien Han <seb@redhat.com>
8 years agoshrink-osd: fix when multiple osds 1885/head
Sébastien Han [Tue, 12 Sep 2017 23:05:34 +0000 (17:05 -0600)]
shrink-osd: fix when multiple osds

The loop was being built properly so we were always getting the last
item as osd host.

Closes: https://bugzilla.redhat.com/show_bug.cgi?id=1490355
Signed-off-by: Sébastien Han <seb@redhat.com>
8 years agoMerge pull request #1880 from ceph/wip-rgw-nfs
Sébastien Han [Wed, 13 Sep 2017 20:20:14 +0000 (14:20 -0600)]
Merge pull request #1880 from ceph/wip-rgw-nfs

nfs: configure RGW FSAL to start up correctly

8 years agotests: delete journal partitions in lvm_setup.yml 1890/head
Andrew Schoen [Wed, 13 Sep 2017 20:02:54 +0000 (15:02 -0500)]
tests: delete journal partitions in lvm_setup.yml

Delete these before creating them incase they are left around in a purge
cluster testing scenario. The purge-cluster.yml playbook does not
currently remove partitions used for journals.

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
8 years agoMerge pull request #1882 from ceph/multi-journal
Sébastien Han [Wed, 13 Sep 2017 17:43:48 +0000 (11:43 -0600)]
Merge pull request #1882 from ceph/multi-journal

osd: drop support for device partition

8 years agoMerge pull request #1888 from ceph/make-srpm-rpmbuild
Sébastien Han [Wed, 13 Sep 2017 17:41:10 +0000 (11:41 -0600)]
Merge pull request #1888 from ceph/make-srpm-rpmbuild

Makefile: remove fedpkg dependency for "srpm"

8 years agoMakefile: remove fedpkg dependency for "srpm" 1888/head
Ken Dreyer [Tue, 12 Sep 2017 22:59:44 +0000 (16:59 -0600)]
Makefile: remove fedpkg dependency for "srpm"

It's not currently possible to install EPEL 7's fedpkg alongside the
centos-packager package in CentOS 7 Extras (for /usr/bin/cbs), because
they each depend on slightly different Koji package versions.

This means that we cannot run "make srpm" on a system where
centos-packager is installed, because that depends on fedpkg.

Remove the dependency on fedpkg for the "srpm" target, and run
rpmbuild -bs directly.

The purpose of this change is to make it easier to automatically
build ceph-ansible SRPMs in Jenkins/cbs.centos.org.

8 years agoMerge pull request #1883 from ceph/quick_refact
Sébastien Han [Wed, 13 Sep 2017 04:11:31 +0000 (22:11 -0600)]
Merge pull request #1883 from ceph/quick_refact

osd: refact include of `activate_osds.yml`

8 years agoMerge pull request #1747 from ceph/add-iscsi
Sébastien Han [Wed, 13 Sep 2017 00:06:50 +0000 (02:06 +0200)]
Merge pull request #1747 from ceph/add-iscsi

resync ceph-iscsi-gw with old upstream

8 years agoresync ceph-iscsi-gw with old upstream 1747/head
Sébastien Han [Fri, 4 Aug 2017 18:18:11 +0000 (20:18 +0200)]
resync ceph-iscsi-gw with old upstream

Taken from https://github.com/pcuzner/ceph-iscsi-ansible/tree/tcmu-fixes

Closes: https://bugzilla.redhat.com/show_bug.cgi?id=1454945 and
https://bugzilla.redhat.com/show_bug.cgi?id=1484083
Signed-off-by: Sébastien Han <seb@redhat.com>
8 years agoMerge pull request #1884 from ceph/mon-container-ip
Sébastien Han [Tue, 12 Sep 2017 23:46:18 +0000 (01:46 +0200)]
Merge pull request #1884 from ceph/mon-container-ip

mon: add support for monitor_address block for containers

8 years agoosd: drop support for device partition 1882/head
Sébastien Han [Tue, 12 Sep 2017 21:31:19 +0000 (15:31 -0600)]
osd: drop support for device partition

We have been struggling with this, it's still broken and breaking other
things too now.

Closes: https://bugzilla.redhat.com/show_bug.cgi?id=1490283
Signed-off-by: Sébastien Han <seb@redhat.com>
8 years agoosd: refact include of `activate_osds.yml` 1883/head
Guillaume Abrioux [Tue, 12 Sep 2017 21:49:11 +0000 (23:49 +0200)]
osd: refact include of `activate_osds.yml`

remove duplicate code.

Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
8 years agomon: add support for monitor_address block for containers 1884/head
Sébastien Han [Tue, 12 Sep 2017 22:28:08 +0000 (16:28 -0600)]
mon: add support for monitor_address block for containers

Signed-off-by: Sébastien Han <seb@redhat.com>
8 years agoMerge pull request #1761 from ceph/split_copy_keys
Sébastien Han [Tue, 12 Sep 2017 22:21:50 +0000 (00:21 +0200)]
Merge pull request #1761 from ceph/split_copy_keys

docker: split the task 'copy ceph configs&keys'

8 years agonfs: configure RGW FSAL to start up correctly 1880/head
Ali Maredia [Mon, 11 Sep 2017 03:34:36 +0000 (23:34 -0400)]
nfs: configure RGW FSAL to start up correctly

- Add RGW keyring to nfs node
- Add RGW section to ganesha.conf
- Add RGW section to ceph.conf onf nfs node

Signed-off-by: Ali Maredia <amaredia@redhat.com>
8 years agotests: no need to create a filesystem on /dev/sdc1 for lvm tests
Andrew Schoen [Tue, 12 Sep 2017 20:14:21 +0000 (15:14 -0500)]
tests: no need to create a filesystem on /dev/sdc1 for lvm tests

The partition only needs created and given a gpt label so that a
PARTUUID will exist on the partition.

This task also makes the purge_lvm_osds scenario fail on the second
deployment after purging.

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
8 years agoMerge pull request #1881 from ceph/followup_1724
Sébastien Han [Tue, 12 Sep 2017 15:43:19 +0000 (17:43 +0200)]
Merge pull request #1881 from ceph/followup_1724

ceph-osd: remove deprecated comment in sample file

8 years agoceph-osd: remove deprecated comment in sample file 1881/head
Guillaume Abrioux [Tue, 12 Sep 2017 14:48:10 +0000 (16:48 +0200)]
ceph-osd: remove deprecated comment in sample file

Since #1724 has been merged, this comment is deprecated

Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
8 years agoDocker: split the task 'copy ceph configs&keys' 1761/head
Guillaume Abrioux [Mon, 11 Sep 2017 08:54:03 +0000 (10:54 +0200)]
Docker: split the task 'copy ceph configs&keys'

All keys are copied to all nodes.
This commit split that task in each roles so keys are copied to their
respective nodes.

Closes: https://bugzilla.redhat.com/show_bug.cgi?id=1488999
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
8 years agoMerge pull request #1863 from ceph/rpm-comments
Sébastien Han [Sat, 9 Sep 2017 10:02:42 +0000 (12:02 +0200)]
Merge pull request #1863 from ceph/rpm-comments

rpm: better comments for file removals

8 years agoMerge pull request #1840 from ceph/simply-containers v3.0.0rc7
Sébastien Han [Sat, 9 Sep 2017 00:35:11 +0000 (02:35 +0200)]
Merge pull request #1840 from ceph/simply-containers

docker: simplify variable declaration

8 years agodocker: simplify variable declaration 1840/head
Sébastien Han [Thu, 31 Aug 2017 10:04:40 +0000 (12:04 +0200)]
docker: simplify variable declaration

Less configuration for the user, the container inherit from the global
variables. No more container specific variables.

Signed-off-by: Sébastien Han <seb@redhat.com>
8 years agoMerge pull request #1878 from ceph/add-rbd-mirror
Sébastien Han [Fri, 8 Sep 2017 23:21:12 +0000 (01:21 +0200)]
Merge pull request #1878 from ceph/add-rbd-mirror

Add rbd mirror

8 years agoci: deploy rbd mirror 1878/head
Sébastien Han [Fri, 8 Sep 2017 07:51:03 +0000 (09:51 +0200)]
ci: deploy rbd mirror

Deploy rbd mirorr in cluster scenario

Signed-off-by: Sébastien Han <seb@redhat.com>
8 years agoMerge pull request #1875 from ceph/switch-nfs
Sébastien Han [Fri, 8 Sep 2017 23:09:30 +0000 (01:09 +0200)]
Merge pull request #1875 from ceph/switch-nfs

switch to container: fix ceph nfs

8 years agoswitch to container: fix ceph nfs 1875/head
Sébastien Han [Fri, 8 Sep 2017 17:28:52 +0000 (19:28 +0200)]
switch to container: fix ceph nfs

The service is nfs-ganesha where ceph-nfs@{{ ansible_hostname }} will be
the name of the container.

Signed-off-by: Sébastien Han <seb@redhat.com>
8 years agoci: remove scenario bluestore_docker_cluster
Sébastien Han [Fri, 8 Sep 2017 08:47:02 +0000 (10:47 +0200)]
ci: remove scenario bluestore_docker_cluster

We don't need to bootstrap a full cluster to bootstrap bluestore. We
have individual scenarios for that.

Signed-off-by: Sébastien Han <seb@redhat.com>
8 years agoMerge pull request #1780 from ceph/wip-rgw-nfs
Sébastien Han [Fri, 8 Sep 2017 17:26:02 +0000 (19:26 +0200)]
Merge pull request #1780 from ceph/wip-rgw-nfs

Wip RGW NFS

8 years agoMerge pull request #1871 from ceph/handler-collocate
Guillaume Abrioux [Fri, 8 Sep 2017 16:15:02 +0000 (18:15 +0200)]
Merge pull request #1871 from ceph/handler-collocate

defaults: do not restart unconfigured (yet) daemons

8 years agoMerge pull request #1874 from ceph/rbd-mirror-mem
Sébastien Han [Fri, 8 Sep 2017 15:50:55 +0000 (17:50 +0200)]
Merge pull request #1874 from ceph/rbd-mirror-mem

ceph-rbd-mirror; docker fix typo

8 years agoceph-rbd-mirror; docker fix typo 1874/head
Sébastien Han [Fri, 8 Sep 2017 15:47:48 +0000 (17:47 +0200)]
ceph-rbd-mirror; docker fix typo

Signed-off-by: Sébastien Han <seb@redhat.com>
8 years agonfs: rename host to have ceph- prefix 1780/head
Ali Maredia [Wed, 6 Sep 2017 16:44:19 +0000 (12:44 -0400)]
nfs: rename host to have ceph- prefix

Signed-off-by: Ali Maredia <amaredia@redhat.com>
8 years agonfs-ganesha: add config overrides section
Ali Maredia [Sun, 3 Sep 2017 15:32:44 +0000 (11:32 -0400)]
nfs-ganesha: add config overrides section

Signed-off-by: Ali Maredia <amaredia@redhat.com>
8 years agoMerge pull request #1870 from Logan2211/omit-default-release
Sébastien Han [Fri, 8 Sep 2017 14:55:03 +0000 (16:55 +0200)]
Merge pull request #1870 from Logan2211/omit-default-release

Omit the apt default_release if it is not needed

8 years agomds: rename mds_socket fact 1871/head
Guillaume Abrioux [Fri, 8 Sep 2017 13:57:58 +0000 (15:57 +0200)]
mds: rename mds_socket fact

Rename this fact to keep consistency with handlers in `ceph-defaults`.

Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
8 years agonfs: add automated testing for nfs-ganesha roles
Ali Maredia [Mon, 21 Aug 2017 17:36:38 +0000 (13:36 -0400)]
nfs: add automated testing for nfs-ganesha roles

Signed-off-by: Ali Maredia <amaredia@redhat.com>
8 years agonfs-ganesha: add dev, stable, and rhcs nfs-ganesha's for ceph-nfs role
Ali Maredia [Mon, 21 Aug 2017 12:38:21 +0000 (08:38 -0400)]
nfs-ganesha: add dev, stable, and rhcs nfs-ganesha's for ceph-nfs role

Signed-off-by: Ali Maredia <amaredia@redhat.com>
8 years agoMerge pull request #1873 from ceph/switch-unit
Sébastien Han [Fri, 8 Sep 2017 11:36:04 +0000 (13:36 +0200)]
Merge pull request #1873 from ceph/switch-unit

switch: make osd collection idempotent

8 years agodefaults: do not restart unconfigured (yet) daemons
Sébastien Han [Thu, 7 Sep 2017 21:33:45 +0000 (23:33 +0200)]
defaults: do not restart unconfigured (yet) daemons

In a collocated scenario, where you might put a rgw, a mds and a mon on
the same node you don't want the handler blindly restart all the daemons
on the node. Indeed some of them might not be configured yet.
Implementing a more precise socket detection, for each daemon type.

Closes: https://bugzilla.redhat.com/show_bug.cgi?id=1488813
Signed-off-by: Sébastien Han <seb@redhat.com>
8 years agoswitch: make osd collection idempotent 1873/head
Sébastien Han [Fri, 8 Sep 2017 09:30:45 +0000 (11:30 +0200)]
switch: make osd collection idempotent

This commits allows us to run
switch-from-non-containerized-to-containerized-ceph-daemons.yml multiple
times.

Closes: https://bugzilla.redhat.com/show_bug.cgi?id=1489353
Signed-off-by: Sébastien Han <seb@redhat.com>
8 years agoMerge pull request #1864 from ceph/switch
Sébastien Han [Fri, 8 Sep 2017 07:39:20 +0000 (09:39 +0200)]
Merge pull request #1864 from ceph/switch

switch-from-non-containerized-to-containerized: fix devices

8 years agoswitch-from-non-containerized-to-containerized: fix devices 1864/head
Sébastien Han [Thu, 7 Sep 2017 07:16:31 +0000 (09:16 +0200)]
switch-from-non-containerized-to-containerized: fix devices

If devices is passed through an extra var this register won't work so
let's only register the var is devices is not defined.

Closes: https://bugzilla.redhat.com/show_bug.cgi?id=1489099
Signed-off-by: Sébastien Han <seb@redhat.com>
8 years agoMerge pull request #1803 from ceph/disk-auto
Sébastien Han [Thu, 7 Sep 2017 20:51:40 +0000 (22:51 +0200)]
Merge pull request #1803 from ceph/disk-auto

ceph-osd: fix autodetection activation

8 years agoOmit the apt default_release if it is not needed 1870/head
Logan V [Thu, 7 Sep 2017 16:50:57 +0000 (11:50 -0500)]
Omit the apt default_release if it is not needed

The apt module will fail to downgrade packages properly when defualt
release is unnecessarily defined.

Closes #1869

8 years agoMerge pull request #1867 from ceph/rhcs-version
Andrew Schoen [Thu, 7 Sep 2017 15:51:48 +0000 (10:51 -0500)]
Merge pull request #1867 from ceph/rhcs-version

tests: set ceph_stable_release to luminous for rhcs

8 years agoceph-osd: fix autodetection activation 1803/head
Sébastien Han [Thu, 24 Aug 2017 09:17:56 +0000 (11:17 +0200)]
ceph-osd: fix autodetection activation

Prior to this patch this activation sequence for autodetection was
always skipped because we were asking to activate on device without
partitions, which doesn't make sense.

We also fix the way we lookup for a device, since the data partition is
always numbered 1, we take the min element of the dict.

Closes: https://github.com/ceph/ceph-ansible/issues/1782
Signed-off-by: Sébastien Han <seb@redhat.com>
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
8 years agoMerge pull request #1842 from ceph/purge
Sébastien Han [Thu, 7 Sep 2017 15:36:34 +0000 (17:36 +0200)]
Merge pull request #1842 from ceph/purge

purge: get lockbox mountpoint and unmount it

8 years agotests: set ceph_stable_release to luminous for rhcs 1867/head
Andrew Schoen [Thu, 7 Sep 2017 14:45:56 +0000 (09:45 -0500)]
tests: set ceph_stable_release to luminous for rhcs

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
8 years agopurge: get lockbox mountpoint and unmount it 1842/head
Sébastien Han [Thu, 31 Aug 2017 16:22:34 +0000 (18:22 +0200)]
purge: get lockbox mountpoint and unmount it

Prior command was avoiding the lockbox mountpoint and the playbook was
failing with:

rmtree failed: [Errno 30] Read-only file system:
'/var/lib/ceph/osd-lockbox/4e9d8052-87c2-4fde-a56c-b8c108a3eefc/key-management-mode'

Signed-off-by: Sébastien Han <seb@redhat.com>
8 years agoMerge pull request #1850 from fultonj/issue/1848
Sébastien Han [Thu, 7 Sep 2017 11:51:11 +0000 (13:51 +0200)]
Merge pull request #1850 from fultonj/issue/1848

Add option to create client keyring file but not import it

8 years agoMerge pull request #1859 from ceph/container-limit
Sébastien Han [Thu, 7 Sep 2017 10:51:34 +0000 (12:51 +0200)]
Merge pull request #1859 from ceph/container-limit

container: introduce resource limitation for containers

8 years agoMerge pull request #1862 from ceph/fail-ansible
Sébastien Han [Thu, 7 Sep 2017 06:44:01 +0000 (08:44 +0200)]
Merge pull request #1862 from ceph/fail-ansible

fail if ansible version < 2.3

8 years agofail if ansible version < 2.3 1862/head
Sébastien Han [Wed, 6 Sep 2017 15:52:49 +0000 (17:52 +0200)]
fail if ansible version < 2.3

We only test and support 2.3.x at the moment.

Closes: https://github.com/ceph/ceph-ansible/issues/1858
Signed-off-by: Sébastien Han <seb@redhat.com>
8 years agoMerge pull request #1812 from ceph/switch-migration-conta
Sébastien Han [Thu, 7 Sep 2017 05:30:34 +0000 (07:30 +0200)]
Merge pull request #1812 from ceph/switch-migration-conta

switch-from-non-containerized-to-containerized: mask unit files

8 years agorpm: better comments for file removals 1863/head
Ken Dreyer [Wed, 6 Sep 2017 17:12:54 +0000 (11:12 -0600)]
rpm: better comments for file removals

As of 54d7a81241eac26d87e2bee513df5efd8866a586, we're doing more than
stripping CoreOS files in the RPM now. Move the comments around to
better match the code that does what the comments describe.

The purpose of this change is to make it easier to read this part of the
RPM spec file.

8 years agotests: force docker variable for switch-to-containers scenario 1812/head
Guillaume Abrioux [Tue, 5 Sep 2017 13:18:01 +0000 (15:18 +0200)]
tests: force docker variable for switch-to-containers scenario

we need to force the value of `docker` variable which is initially set
to `false` since it's a migration from non-containerized to
containerized cluster.

Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
8 years agoAdd option to create client keyring file but not import it 1850/head
John Fulton [Sat, 2 Sep 2017 01:30:04 +0000 (01:30 +0000)]
Add option to create client keyring file but not import it

Add new boolean parameter for client config create_key_file_only
with a default of false. When create_key_file_only is true, the
client tasks to connect to the external ceph cluster to verify
the key `ceph auth import` the key are skipped.

Fixes: #1848
8 years agocontainer: introduce resource limitation for containers 1859/head
Sébastien Han [Wed, 6 Sep 2017 08:53:30 +0000 (10:53 +0200)]
container: introduce resource limitation for containers

This can be controlled via 2 options:

* ceph_$DAEMON_docker_memory_limit
* ceph_$DAEMON_docker_cpu_limit

All daemons default to 1GB for memory and 1 CPU by default.
Recommendations from:
https://access.redhat.com/documentation/en-us/red_hat_ceph_storage/2/html/red_hat_ceph_storage_hardware_guide/minimum_recommendations

Signed-off-by: Sébastien Han <seb@redhat.com>
8 years agoMerge pull request #1860 from ceph/ssh-timeout
Guillaume Abrioux [Wed, 6 Sep 2017 09:36:17 +0000 (11:36 +0200)]
Merge pull request #1860 from ceph/ssh-timeout

ansible.cfg: increase ssh timeout