Commit Graph

1380 Commits (fcb0a01e287a20fa3add7ae9544076a066a9f736)

Author SHA1 Message Date
Sébastien Han fb02b1d9d3 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>
2017-09-14 11:06:44 -06:00
Sébastien Han 6f0b1fe803 rgw: remove old variables
Since the only support civetweb these variables are obsolete.

Signed-off-by: Sébastien Han <seb@redhat.com>
2017-09-14 09:42:50 -06:00
Sébastien Han ef8d37dd0d Merge pull request #1800 from ceph/wip-osd-start-fix
ceph-osd: Fix osd start sequence
2017-09-13 17:20:10 -06:00
Sébastien Han 2f51f0de28 Merge pull request #1880 from ceph/wip-rgw-nfs
nfs: configure RGW FSAL to start up correctly
2017-09-13 14:20:14 -06:00
Sébastien Han f67b47d056 Merge pull request #1882 from ceph/multi-journal
osd: drop support for device partition
2017-09-13 11:43:48 -06:00
Sébastien Han ac62437609 Merge pull request #1883 from ceph/quick_refact
osd: refact include of `activate_osds.yml`
2017-09-12 22:11:31 -06:00
Sébastien Han c3866fc4bd Merge pull request #1747 from ceph/add-iscsi
resync ceph-iscsi-gw with old upstream
2017-09-13 02:06:50 +02:00
Sébastien Han aa364264cd 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>
2017-09-12 18:06:10 -06:00
Sébastien Han 2a1b8a1997 Merge pull request #1884 from ceph/mon-container-ip
mon: add support for monitor_address block for containers
2017-09-13 01:46:18 +02:00
Sébastien Han fdf924401f 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>
2017-09-12 17:42:07 -06:00
Guillaume Abrioux 49ad8528e5 osd: refact include of `activate_osds.yml`
remove duplicate code.

Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
2017-09-12 16:53:11 -06:00
Sébastien Han 02ba65dbbe mon: add support for monitor_address block for containers
Signed-off-by: Sébastien Han <seb@redhat.com>
2017-09-12 16:28:08 -06:00
Sébastien Han 6b8ed0440e Merge pull request #1761 from ceph/split_copy_keys
docker: split the task 'copy ceph configs&keys'
2017-09-13 00:21:50 +02:00
Ali Maredia 52efe92a87 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>
2017-09-12 16:27:16 -04:00
Guillaume Abrioux 20946f7220 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>
2017-09-12 16:48:10 +02:00
Guillaume Abrioux 0f506f4f0a 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>
2017-09-11 21:14:13 +02:00
Sébastien Han 2ea7f287fa 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>
2017-09-09 01:22:06 +02:00
Sébastien Han 4767eaaab3 Merge pull request #1878 from ceph/add-rbd-mirror
Add rbd mirror
2017-09-09 01:21:12 +02:00
Sébastien Han 7054615551 ci: deploy rbd mirror
Deploy rbd mirorr in cluster scenario

Signed-off-by: Sébastien Han <seb@redhat.com>
2017-09-09 01:17:10 +02:00
Sébastien Han 477f86e305 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>
2017-09-08 22:43:50 +02:00
Sébastien Han d46d453b83 Merge pull request #1780 from ceph/wip-rgw-nfs
Wip RGW NFS
2017-09-08 19:26:02 +02:00
Guillaume Abrioux b59e9cc732 Merge pull request #1871 from ceph/handler-collocate
defaults: do not restart unconfigured (yet) daemons
2017-09-08 18:15:02 +02:00
Sébastien Han a05c58ba37 Merge pull request #1874 from ceph/rbd-mirror-mem
ceph-rbd-mirror; docker fix typo
2017-09-08 17:50:55 +02:00
Sébastien Han 7a93d88025 ceph-rbd-mirror; docker fix typo
Signed-off-by: Sébastien Han <seb@redhat.com>
2017-09-08 17:47:48 +02:00
Ali Maredia f8171e8b4a nfs: rename host to have ceph- prefix
Signed-off-by: Ali Maredia <amaredia@redhat.com>
2017-09-08 11:38:05 -04:00
Ali Maredia f3e2235b3a nfs-ganesha: add config overrides section
Signed-off-by: Ali Maredia <amaredia@redhat.com>
2017-09-08 11:37:58 -04:00
Sébastien Han d53f55e807 Merge pull request #1870 from Logan2211/omit-default-release
Omit the apt default_release if it is not needed
2017-09-08 16:55:03 +02:00
Guillaume Abrioux 44fd928e23 mds: rename mds_socket fact
Rename this fact to keep consistency with handlers in `ceph-defaults`.

Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
2017-09-08 15:57:58 +02:00
Ali Maredia 55724c6e93 nfs-ganesha: add dev, stable, and rhcs nfs-ganesha's for ceph-nfs role
Signed-off-by: Ali Maredia <amaredia@redhat.com>
2017-09-08 09:13:20 -04:00
Sébastien Han 12f6e53090 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>
2017-09-08 12:02:37 +02:00
Logan V d8cb62c981 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
2017-09-07 11:50:57 -05:00
Sébastien Han 3753e6cfa7 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>
2017-09-07 17:47:37 +02:00
Sébastien Han 27b3f9a7d4 Merge pull request #1850 from fultonj/issue/1848
Add option to create client keyring file but not import it
2017-09-07 13:51:11 +02:00
Sébastien Han cf88c136f5 Merge pull request #1859 from ceph/container-limit
container: introduce resource limitation for containers
2017-09-07 12:51:34 +02:00
Sébastien Han d2032c92af Merge pull request #1862 from ceph/fail-ansible
fail if ansible version < 2.3
2017-09-07 08:44:01 +02:00
Sébastien Han fc3300ea4f 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>
2017-09-07 07:53:17 +02:00
John Fulton a57f61efd9 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
2017-09-06 13:56:06 +00:00
Sébastien Han 2fa151b9e8 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>
2017-09-06 14:52:21 +02:00
Sébastien Han b7db600caa switch-from-non-containerized-to-containerized: mask unit files
We must mask the image so we are sure that even if the system reboots
then the OSDs won't start.

Also remove Ceph udev rules if found on the system prior to deploy
containers. If we don't do this we are exposed to conflicts between udev
rules and sytemd unit files.

Also add the CI will now test the migration from a non-containerized cluster to a
containerized cluster.

Signed-off-by: Sébastien Han <seb@redhat.com>
2017-09-05 15:20:31 +02:00
Sébastien Han b6c1a0c68f Merge pull request #1853 from ceph/fix-prepare
ceph-osd: do not re-prepare if already prepared
2017-09-05 13:59:40 +02:00
Sébastien Han 5ed1a91aeb Merge pull request #1819 from ceph/no-container-log
ceph-docker-common: do not log inside the container
2017-09-05 11:47:11 +02:00
Sébastien Han 1dd976d28e ceph-osd: do not re-prepare if alreadyy prepared
I forgot to re-add the partition check while refactoring the osd

Signed-off-by: Sébastien Han <seb@redhat.com>
2017-09-05 09:51:57 +02:00
Sébastien Han 23a0c26c4f client: do not copy admin key by default
Signed-off-by: Sébastien Han <seb@redhat.com>
2017-09-02 00:54:17 +02:00
Sébastien Han 58f664fd17 ceph-rgw: fix systemd unit layout
Signed-off-by: Sébastien Han <seb@redhat.com>
2017-09-01 19:02:48 +02:00
Sébastien Han 967e875fd0 Merge pull request #1827 from andymcc/rgw_systemd_fix
Fix RGW systemd directory
2017-09-01 18:12:23 +02:00
Alfredo Deza 98d107cebb common do not filter by distro when dev is set for figuring out ceph_release
Signed-off-by: Alfredo Deza <adeza@redhat.com>
2017-08-31 16:18:08 -04:00
Sébastien Han 673938ec96 Merge pull request #1839 from ceph/colonwq-update-docker-rgw-exec
Update ceph_rgw_docker_extra_env to add bind ip
2017-08-31 19:47:16 +02:00
Sébastien Han ea9b6395cb Merge pull request #1838 from ceph/rgw-units
Rgw units
2017-08-31 19:38:23 +02:00
Andrew Schoen 29df79e54e Merge pull request #1841 from ceph/lvm-partitions
lvm-osds: test with a partition and an lv as journals
2017-08-31 12:09:19 -05:00
Sébastien Han 3dd47a45cb ceph-defaults: fix handlers for mds and rgw
The way we handle the restart for both mds and rgw is not ideal, it will
try to restart the daemon on the host that don't run the daemon,
resulting in a service file being created (see bug description).

Now we restart each daemon precisely and in a serialized fashion.

Note: the current implementation does NOT support multiple mds or rgw on
the same node.

Closes: https://bugzilla.redhat.com/show_bug.cgi?id=1469781
Signed-off-by: Sébastien Han <seb@redhat.com>
2017-08-31 19:02:21 +02:00