Commit Graph

404 Commits (6c3db809aef730acd7c53ec3260c2a4cfb6a2350)

Author SHA1 Message Date
Leseb cc894d4044 Merge pull request #800 from ceph/BZ-1337320
ceph-common: on rh storage on debian add the repositories and keys (ISO)
2016-05-20 14:18:07 +02:00
Leseb 5984888a2f Merge pull request #797 from ceph/rgw-debian-rhcs
install ceph-mds and radosgw on debian systems when using red hat ceph storage
2016-05-20 14:07:02 +02:00
Alfredo Deza ed22cfa213 ceph-common: redhat repo template: remove the Calamari reference, no longer a distinct repo
Signed-off-by: Alfredo Deza <adeza@redhat.com>
2016-05-19 16:17:02 -04:00
Alfredo Deza c78313d19f ceph-common: on rh storage on debian do not add the rh key by default
Signed-off-by: Alfredo Deza <adeza@redhat.com>
2016-05-19 15:16:18 -04:00
Alfredo Deza 329ac01d49 ceph-common: on rh storage on debian add the repositories and keys
Signed-off-by: Alfredo Deza <adeza@redhat.com>
2016-05-19 14:20:47 -04:00
Leseb 7f85b0d5ff Merge pull request #796 from ceph/fix-rgw-frontends
fix the radosgw_civetweb_bind_ip default value
2016-05-19 16:14:32 +01:00
Andrew Schoen eb92cb57e4 install ceph-mds on debian systems when using red hat ceph storage
Signed-off-by: Andrew Schoen <aschoen@redhat.com>
2016-05-19 09:56:06 -05:00
Andrew Schoen 7157afc7ce install radosgw on debian systems when using red hat ceph storage
Signed-off-by: Andrew Schoen <aschoen@redhat.com>
2016-05-19 09:52:18 -05:00
Andrew Schoen a6a946d86b fix the radosgw_civetweb_bind_ip default value
We want just the address for ansible_default_ipv4, not the entire
dictionary in ceph.conf.

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
2016-05-18 15:53:47 -05:00
Alfredo Deza 7b8c16b4e9 ceph-common: remove RedHat-only constraint for RH ISO install
Signed-off-by: Alfredo Deza <adeza@redhat.com>
2016-05-18 16:40:28 -04:00
Sébastien Han 1d5d438e5d ceph-common: list impacted versions
For the ceph option default path

Signed-off-by: Sébastien Han <seb@redhat.com>
2016-05-18 15:25:49 +01:00
Sébastien Han e071b773d8 ceph-common: add explanation message
Signed-off-by: Sébastien Han <seb@redhat.com>
2016-05-16 18:55:54 +02:00
Sébastien Han e7aae27801 ceph-common: fix ceph options default path again
closes: #788

Signed-off-by: Sébastien Han <seb@redhat.com>
2016-05-16 18:37:28 +02:00
Andrew Schoen 85fb03fc99 fix the /etc/default/ceph path on debian when setting cluster name
In ceph version 10.2.1 a bug was fixed that corrects the path:

791eba81a5

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
2016-05-13 15:22:24 -05:00
Leseb 82a249e742 Merge pull request #771 from ceph/client-role
ceph-client: introduce new client role
2016-05-11 14:40:15 +02:00
Sébastien Han 40df8fb814 ceph-client: introduce new client role
By default, this roles will create a ceph config file and get the admin
key. You can optionnally add other users, keys and pools for your tests.

Closes: #769

Signed-off-by: Sébastien Han <seb@redhat.com>
2016-05-11 14:36:35 +02:00
Samuel Matzek 9fec57a583 Add Debian ppc64le support
Add support to allow ceph-ansible to install and
configure Ceph on Debian on the ppc64le architecture.

Canonical has ppc64le Debian packages in Ubuntu distros
and on Ubuntu Cloud Archive.  Both of which can be installed
and configured using the 'distro' or 'uca' options in
ceph-ansible when this patch is used.

Signed-off-by: Samuel Matzek <smatzek@us.ibm.com>
2016-05-10 14:26:30 -05:00
Leseb 52b2f1cb85 Merge pull request #694 from ceph/purge-ceph-conf-options
ceph-common: purge ceph.conf file
2016-05-10 18:24:07 +02:00
Sébastien Han 47860a887f ceph-common: purge ceph.conf file
Since ##461 we have been having the ability to override ceph default
options. Previously we had to add a new line in the template and then
another variable as well. Doing a PR for one option was such a pain. As
a result, we now have tons of options that we need to maintain across
all the ceph version, yet another painful thing to do.
This commit removes all the ceph options so they are handled by ceph
directly. If you want to add a new option, feel free to to use the
`ceph_conf_overrides` variable of your `group_vars/all`.

Risks, for those who have been managing their ceph using ceph-ansible
this is not a trivial change as it will trigger a change in your
`ceph.conf` and then restart all your ceph services. Moreover if you did
some specific tweaks as well, prior to run ansible you should update the
`ceph_conf_overrides` variable to reflect your previous changes.

To avoid service restart, you need to know a bit of ansible for this,
but generally the idea would be to run ansible on a dummy host to
generate the ceph.conf, then scp this file to all your ceph hosts and
you should be good.

Closes: #693

Signed-off-by: Sébastien Han <seb@redhat.com>
2016-05-10 16:51:42 +02:00
Stephan Hohn 4dad21c481 Increase system stabillity by re-enabling swap and setting vm.swappiness to 10 2016-05-09 22:05:01 +02:00
Sam Yaple 14d31c7bcd Add in missing is_before_infernalis statement 2016-05-09 15:26:37 +00:00
Sam Yaple 1d5406c4a3 Fixed a few missed when conditionals from rebase 2016-05-09 14:17:14 +00:00
Sam Yaple 069c93a238 Unify formatting of when conditional
This is purely a refactor. Converts when 'and' conditionals into lists
rather than multiline strings. This does not work for nested
conditionals, but those can be formated with indents.

Moves one line when statements onto the same line as the when command
itself.

A small logic bug was found in ceph-osd/tasks/check_devices.yml which
which was also fixed.

Signed-off-by: Sam Yaple <sam@yaple.net>
2016-05-09 14:08:33 +00:00
Leseb 30f7208334 Merge pull request #748 from ceph/use-jewel
ceph: use jewel release
2016-05-09 01:13:54 +02:00
Leseb 1b8d021ef0 Merge pull request #764 from ceph/systemd
Several fixes
2016-05-09 01:00:02 +02:00
Sébastien Han b3971354bf ceph-mds: fix the handler
Moving the mds_name fact in the common role so the MDS can properly
be restarted.

Signed-off-by: Sébastien Han <seb@redhat.com>
2016-05-09 00:39:03 +02:00
Sébastien Han c4f3ab5649 ceph-common: install ceph-mds package
We kinda ommitted this package, now Jewel is out and the metadata server
is considered as stable.

Signed-off-by: Sébastien Han <seb@redhat.com>
2016-05-07 02:37:09 +02:00
Andrew Schoen 5139ed0592 strip any newlines from the init_system fact
Signed-off-by: Andrew Schoen <aschoen@redhat.com>
2016-05-06 14:47:45 -05:00
Andrew Schoen cb70d49751 Use the use_systemd fact when starting mdss
Signed-off-by: Andrew Schoen <aschoen@redhat.com>
2016-05-06 14:47:45 -05:00
Andrew Schoen 9eed0ea4e7 Adds an use_systemd fact
This adds a helper fact that uses the ``init_system`` fact to determine if
we should be using systemd or not when controlling services.

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
2016-05-05 11:53:08 -05:00
Leseb 480cb68892 Merge pull request #755 from mattt416/fix_issue_741
Set init_system fact and reference in ceph-osd role
2016-05-04 16:39:41 +02:00
Matt Thompson ed2b7757d4 Set init_system fact and reference in ceph-osd role
The ceph-osd role currently uses ansible_service_mgr, which is a fact
only available on ansible 2.x and greater.  This commit sets a similar
fact called init_system which will store the contents of /proc/1/comm
(systemd, init, etc.) and then references it ceph-osd instead.

Closes #741
2016-05-04 12:05:31 +01:00
Samuel Matzek ae239aaf09 Add support for install from Ubuntu Cloud Archive
This adds support to allow the install of Ceph from the
Ubuntu Cloud Archive.  The Ubuntu Cloud Archive provides newer
release of Ceph than the normal Ubuntu distro repository.

Signed-off-by: Samuel Matzek <smatzek@us.ibm.com>
2016-05-03 10:51:39 -05:00
Sébastien Han 94ce2be758 ceph: use jewel release
Also fix the repo url.

Signed-off-by: Sébastien Han <seb@redhat.com>
2016-05-03 14:02:44 +02:00
Sébastien Han 6513cdb313 ceph-common: install ceph-mon/osd for ceph dev
Since developement versions of Ceph are after infernalis a package split
happened. So basically ceph-mon, ceph-osd, ceph-mds need to be
installed.

Signed-off-by: Sébastien Han <seb@redhat.com>
2016-05-02 14:18:33 +02:00
Andrew Schoen f81140d21f Adds a use_server_package_split var
This will allow a user to conditionally install the ceph package on rpm
based systems. Installing this package is not required or wanted in
versions passed infernalis.

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
2016-04-28 09:15:05 -05:00
Alfredo Deza 3af17a3137 ceph-common: port main handler to use new ceph version facts
Signed-off-by: Alfredo Deza <adeza@redhat.com>
2016-04-27 13:46:52 -04:00
Alfredo Deza df31745911 ceph-common: port main task to use new ceph version facts
Signed-off-by: Alfredo Deza <adeza@redhat.com>
2016-04-27 10:21:16 -04:00
Alfredo Deza be8ae40bfb create is_before_RELEASE vars as well
Signed-off-by: Alfredo Deza <adeza@redhat.com>
2016-04-27 10:21:16 -04:00
Alfredo Deza 8f35f1fb9b ceph-common: include the facts task always
Signed-off-by: Alfredo Deza <adeza@redhat.com>
2016-04-27 10:21:16 -04:00
Alfredo Deza 25cc1ec1d3 ceph-common: create a facts task to define ceph versions
Signed-off-by: Alfredo Deza <adeza@redhat.com>
2016-04-27 10:21:16 -04:00
Leseb 7bd44d0563 Merge pull request #721 from ceph/fix-admin-socket-warning
change defaults for rbd_client_directories and rbd_client_admin_socket_path
2016-04-23 01:00:23 +02:00
Andrew Schoen 79a6e05228 default rbd_client_log_path to /var/log/ceph
Signed-off-by: Andrew Schoen <aschoen@redhat.com>
2016-04-22 12:39:16 -05:00
Andrew Schoen b649c4229a default rbd_client_admin_socket_path to /var/run/ceph
Signed-off-by: Andrew Schoen <aschoen@redhat.com>
2016-04-22 10:59:00 -05:00
Andrew Schoen f297af23c5 default rbd_client_directories to true
Signed-off-by: Andrew Schoen <aschoen@redhat.com>
2016-04-22 10:56:30 -05:00
Sébastien Han 4fa4154b4d ceph-rgw: ability to bind civetweb on an IP
Introducing a new config option: `radosgw_civetweb_bind_ip` which points
to the `ansible_default_ipv4` by default. You can override this
variable. Use ansible facts to put a proper value.

Signed-off-by: Sébastien Han <seb@redhat.com>
2016-04-22 17:47:57 +02:00
Andrew Schoen 95d80ad3d3 correctly set mon addr when monitor_interface is defined in group_vars
This fixes the ceph.conf template so that it will look for an inventory
defined value for monitor_interface or for monitor_interface defined in
a group_vars file.

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
2016-04-21 09:09:16 -05:00
Leseb f950afef92 Merge pull request #664 from jsaintrocc/master
Add infernalis to list of stable releases comment
2016-04-19 10:09:48 +02:00
Leseb f51548580f Merge pull request #713 from stpierre/fix-rbd-client-dirs
Fix creation of RBD client directories
2016-04-13 22:25:59 +02:00
Chris St. Pierre de22ae0a06 Fix creation of RBD client directories
Need to use a template, not a bare variable name.
2016-04-13 15:10:44 -05:00
Andrew Schoen a2b8ef8c1d Check to see if hostvars[host]['monitor_interface'] is defined
This fixes https://github.com/ceph/ceph-ansible/issues/711

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
2016-04-13 13:34:51 -05:00
James Saint-Rossy fc270bb772 fixed handling of new ceph_version 2016-04-12 21:26:59 -04:00
James Saint-Rossy ee1e470757 moved string parsing of ceph version into ansbile for better error handling 2016-04-12 21:08:42 -04:00
Leseb 74e3e04dcd Merge pull request #696 from stpierre/dedup-rbd-client-dirs
Deduplicate RBD client directory creation
2016-04-12 12:03:56 +02:00
James Saint-Rossy 3f9bd5c1b7 Removed infernalis workaround because ceph_stable_releases now includes it by default 2016-04-09 17:21:36 -04:00
James Saint-Rossy 4e7b93d725 Rebasing fork to account for upstream commits 2016-04-09 14:40:00 -04:00
Andrew Schoen 5031234d66 look at monitor_interface in hostvars when choosing interface or address
This fixes a bug where monitor_interface might be set in your inventory
file and not by using group_vars or --extra-vars causing the template to
use the default address of 0.0.0.0 instead of the defined
monitor_interface.

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
2016-04-08 14:23:12 -05:00
Chris St. Pierre 1efe62252c Deduplicate RBD client directory creation
Instead of creating the RBD client socket path three different places
in three different ways, this creates it once. Ceph on OpenStack users
have the option to customize the permissions of the RBD client
directories.

Fixes #687
2016-04-08 13:50:00 -05:00
Leseb f783adcc9e Merge pull request #678 from ceph/rbd-mirror
ceph-rbd-mirror: add new role
2016-04-08 12:12:30 +02:00
Sébastien Han 2d21f741f5 ceph-rbd-mirror: add new role
Signed-off-by: Sébastien Han <seb@redhat.com>
2016-04-08 12:12:15 +02:00
Leseb 0047e83dad Merge pull request #684 from bengland2/when-hammer-try2
fix when clauses for installation of ceph RPM
2016-04-06 19:38:05 +02:00
Ben England ee9093d570 Sebastien's suggestion - add infernalis to list of releases with no separate ceph-mon or ceph-osd 2016-04-06 13:14:24 -04:00
Leseb b79e7287cc Merge pull request #685 from bengland2/create-sysconfig-ceph
create /etc/sysconfig/ceph if it doesn't already exist
2016-04-06 13:21:00 +02:00
Ben England dc2fc3665e create /etc/sysconfig/ceph if it doesn't already exist 2016-04-05 21:33:05 -04:00
Ben England b199bdcfab fix when clauses for installation of ceph RPM 2016-04-05 21:11:37 -04:00
Sébastien Han a8df31cb14 ceph-common: let ceph handle pool_default_min_size
We now set the value to 0, which means no specific default; ceph will
use size-size/2

Signed-off-by: Sébastien Han <seb@redhat.com>
2016-04-05 17:01:49 +02:00
Sébastien Han 84e57724e3 ceph-common: change the pool_default_size to 3
3 has been the default since Firefly I believe so we should stick with
this value.

Signed-off-by: Sébastien Han <seb@redhat.com>
2016-04-05 16:22:29 +02:00
James Saint-Rossy b55df73061 Cleanup is_ceph_infernalis logic 2016-04-04 18:19:49 -04:00
James Saint-Rossy ae96854cda Fixed missing release, infernalis, in list of stable releases 2016-03-30 15:55:51 -04:00
Sébastien Han 450feaac0a ceph: implement cluster name support
we now have the ability to enable the `cluster` variable with a specific
value that will determine the name of the cluster.

Signed-off-by: Sébastien Han <seb@redhat.com>
2016-03-30 18:08:38 +02:00
Leseb 629108fbfa Merge pull request #657 from ksingh7/fix_mon_containerized_deployment_with_kv
ceph-common: Newline separation in ceph.conf.j2
2016-03-29 00:56:36 +02:00
ksingh7 861dd2c3f1 ceph-common: Newline separation for mon_containerized_deployment_with_kv variable in ceph.conf.j2
ceph.conf.j2 template requires a new line between mon_containerized_deployment_with_kv and fsid variables
With this commit , i have added a new line for better readablity
2016-03-29 01:34:08 +03:00
Leseb e539645fba Merge pull request #653 from ksingh7/fix_mon_containerized_deployment_with_kv
ceph-common: Fix for issue no. 652
2016-03-28 20:27:39 +02:00
runsisi a2d1d99ec3 ceph-common: use variable rgw_group_name instead of hard coded group
name 'rgws' when checking rgw port.

Signed-off-by: runsisi <runsisi@hust.edu.cn>
2016-03-28 18:59:09 +08:00
ksingh7 06e92923fa ceph-common: Fix for issue no. 652
ceph.conf file generation task in ceph-common role was getting failed
because it ansible cant find defination of varriable mon_containerized_deployment_with_kv
This fix declare mon_containerized_deployment_with_kv under ceph-common/defaults/main.yml which fixes this issue

Signed-off-by: ksingh7 <karan.singh731987@gmail.com>
2016-03-26 19:47:40 +02:00
Leseb f27fada28a Merge pull request #649 from ceph/bluestore-ceph-disk
ceph-osd: add support for bluestore
2016-03-25 17:00:41 +01:00
Leseb e40e0f8950 Merge pull request #650 from stpierre/rbd-client-directory-owner
Make RBD client directory owner/group configurable
2016-03-25 16:41:31 +01:00
Sébastien Han 9098eb0a8a ceph-common: pass journal check if using bluestore
bluestore does not need any journal so we skip this test if
osd_objectstore = bluestore

Signed-off-by: Sébastien Han <seb@redhat.com>
2016-03-25 15:49:39 +01:00
Chris St. Pierre 23e40e3113 Make RBD client directory owner/group configurable
Some versions (?) of libvirt provide a 'libvirt' group instead of
'libvirtd'. (Observed with libvirt-daemon-1.2.17-13.el7_2.2.x86_64.)
This makes the RBD client directory owner and group configurable to
allow for this.
2016-03-25 08:51:20 -05:00
Sébastien Han b0f56590e0 docker: fix tons of issues
Signed-off-by: Sébastien Han <seb@redhat.com>
2016-03-24 17:55:21 +01:00
pprokop 77bd4aff37 Adding missing space 2016-03-24 16:27:56 +01:00
pprokop 832c3bedba Adding comments 2016-03-24 16:27:21 +01:00
pprokop 72fa8da1dc Removing fsid with kv backend deployment 2016-03-24 16:26:28 +01:00
pprokop 6fea8b17b6 Adding missing fsid 2016-03-24 16:26:28 +01:00
pprokop 42e748a514 Adding dockerized ceph-mon deployment with kv backend 2016-03-24 16:26:28 +01:00
Huamin Chen f88eff37d7 initial steps to provision CentOS Atomic host
Signed-off-by: Huamin Chen <hchen@redhat.com>
2016-03-24 16:13:15 +01:00
Andrew Schoen 73a60bd5e0 docs for usage of monitor_interface and monitor_address
Signed-off-by: Andrew Schoen <aschoen@redhat.com>
2016-03-23 07:30:01 -05:00
Andrew Schoen 959d525f81 Provide the ability to provide monitor_address instead of an interface
This would allow users who don't know what interface to provide to
give an IP address to use for the monitor instead.

Note: the includes are needed in ceph.conf.j2 because without them
jinja2 can not properly evaluate the template and will complain about a
missing 'ansible_interface' variable. The includes allow the template to
be evaluated correctly and then the correct include will be used during
render time.

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
2016-03-23 07:30:01 -05:00
Leseb db746e03cb Merge pull request #627 from ceph/fix-ceph-install
ceph-common: fix ceph install for centos distros
2016-03-18 14:02:22 +01:00
Leseb c8422e29ff Merge pull request #633 from ceph/centos-libselinux-python
add libselinux-python as dependency pkg for centos
2016-03-17 17:39:55 +01:00
Sébastien Han 71a0f27f39 add libselinux-python as dependency pkg for centos
closes: #631

Signed-off-by: Sébastien Han <seb@redhat.com>
2016-03-17 17:39:01 +01:00
Sébastien Han 1764b5337f ceph-common: fix ceph install for centos distros
closes: #616

Signed-off-by: Sébastien Han <seb@redhat.com>
2016-03-17 11:55:33 +01:00
Leseb 6e60fb5c0d Merge pull request #610 from fcharlier/install_ceph_mds_only_if_required
Install ceph-mds on Debian platforms only if mds_group_name is set
2016-03-16 14:04:20 +01:00
Andrew Schoen 565e2b280f adds the rh storage apt-key for jewel on ubuntu
Signed-off-by: Andrew Schoen <aschoen@redhat.com>
2016-03-08 16:08:53 -06:00
Andrew Schoen e09bb11dc9 the ceph package does not exist in jewel rh storage
Signed-off-by: Andrew Schoen <aschoen@redhat.com>
2016-03-08 16:06:56 -06:00
Andrew Schoen 3035f45c51 adds the ability to install red hat ceph storage packages on Ubuntu
This will only work for jewel or later releases.

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
2016-03-08 15:23:32 -06:00
Andrew Schoen 60746d9327 subscription manager tasks should not run on debian systems
Signed-off-by: Andrew Schoen <aschoen@redhat.com>
2016-03-08 14:56:23 -06:00
François Charlier 59f151f374 Install ceph-mds on Debian platforms only if mds_group_name is set
Fixes #547
2016-03-08 19:08:24 +01:00
Andrew Schoen 9f1b437e9d find the rhcs version right before setting is_ceph_infernalis
Where it was located before meant it might be skipped if you don't run
tasks with the package-install tag. This fixes the situation where you
want to configure an rhcs node, but do not want to do any package
installs.

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
2016-03-07 14:50:05 -06:00
Andrew Schoen ad6ebbcb32 make sure ceph_origin != "upstream" when using rhcs
When installing RHCS there is an option to install from distro provided
packages, this commit modifies the check to allow that to happen.

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
2016-03-07 14:47:09 -06:00