When devices is not defined because you want to use the 'lvm'
osd_scenario but you've made a mistake selecting that scenario these
tasks should not fail.
Signed-off-by: Andrew Schoen <aschoen@redhat.com>
Extra space in systemctl list-units can cause restart_osd_daemon.sh to
fail
It looks like if you have more services enabled in the node space
between "loaded" and "active" get more space as compared to one space
given in command the command[1].
Closes: https://bugzilla.redhat.com/show_bug.cgi?id=1573317
Signed-off-by: Sébastien Han <seb@redhat.com>
Check whether a mgr module is supposed to be disabled before disabling
it and whether it is already enabled before enabling it.
Signed-off-by: Michael Vollman <michael.b.vollman@gmail.com>
We can simply reference the template name since it exists within the
role that we are calling. We don't need to check the ANSIBLE_ROLE_PATH
or playbooks directory for the file.
To make the package installation more efficient we should install
packages as a list rather than as individual tasks or using a
"with_items" loop. The package managers can handle a list passed to them
to install in one go.
We can use a specified list and substitute any packages that are not to
be installed with the ceph-common package, which is installed on every
package install, then apply the unique filter to the package install
list.
There is no need to stat for created mgr keyrings since they are created
anyway when deploying a ceph cluster > jewel. In case of a jewel
deployment we won't enter that block.
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
This file is a leftover from PR ceph/ceph-ansible#2516
It is not used anymore so it can be removed.
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
Until all the mons haven't been updated to Luminous, there is no way to
create a key. So we should do the key creation in the mon role only if
we are not part of an update.
If we are then the key creation is done after the mons upgrade to
Luminous.
Closes: https://bugzilla.redhat.com/show_bug.cgi?id=1574995
Signed-off-by: Sébastien Han <seb@redhat.com>
trying to mask target when `/etc/systemd/system/target.service` doesn't
exist seems to be a bug.
There is no need to mask a unit file which doesn't exist.
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
The order of fs.aio-max-nr (which is hard-coded to 1048576) means that
if you set fs.aio-max-nr in os_tuning_params it will effectively be
ignored for bluestore scenarios.
To resolve this we should move the setting of fs.aio-max-nr above the
setting of os_tuning_params, in this way the operator can define the
value of fs.aio-max-nr to be something other than 1048576 if they want
to.
Additionally, we can make the sysctl settings happen in 1 task rather
than multiple.
trying to set the default value for pg_num to
`hostvars[groups[mon_group_name][0]]['osd_pool_default_pg_num'])` will
break in case of external client nodes deployment.
the `pg_num` attribute should be mandatory and be tested in future
`ceph-validate` role.
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
On containerized deployment,
when upgrading from jewel to luminous, mgr keyring creation fails because the
command to create mgr keyring is executed on a container that is still
running jewel since the container is restarted later to run the new
image, therefore, it fails with bad entity error.
To get around this situation, we can delegate the command to create
these keyrings on the first monitor when we are running the playbook on the last monitor.
That way we ensure we will issue the command on a container that has
been well restarted with the new image.
Closes: https://bugzilla.redhat.com/show_bug.cgi?id=1574995
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
The Debian and SuSE installs for nfs-ganesha on the non-rhcs repository
requires you to allow_unauthenticated for Debian, and disable_gpg_check
for SuSE. The nfs-ganesha-rgw package already does this, but the
nfs-ganesha-ceph package will fail to install because of this same
issue.
This PR moves the installations to happen when the appropriate flags are
set to True (nfs_obj_gw & nfs_file_gw), but does it per distro (one for
SuSE and one for Debian) so that the appropriate flag can be passed to
ignore the GPG check.
When 'ceph_nfs_disable_caching' is set to True, disable attribute
caching done by Ganesha for all Ganesha exports.
Signed-off-by: Ramana Raja <rraja@redhat.com>
If we are in a middle of an update we want to get the new package
version being installed so the task that copies the repo files should
not be skipped.
Closes: https://bugzilla.redhat.com/show_bug.cgi?id=1572032
Signed-off-by: Sébastien Han <seb@redhat.com>
The apt-cache update can fail due to transient issues related to the
action being a network operation. To reduce the impact of these
transient failures this patch adds a retry to the update_cache task.
However, the apt_repository tasks which would perform an apt_update
won't retry the apt_update on a failure in the same way, as such this PR
moves the apt_update into an individual task, once per role.
Finally, the apt_repository tasks no longer have a changed_when: false,
and the apt_cache update is only performed once per role, if the
repositories change. Otherwise the cache is updated on the "apt" install
tasks if the cache_timeout has been reached.
the value in `docker_exec_client_cmd` doesn't allow to check for
existing pools because it's set with a wrong value for the entrypoint
that is going to be used.
It means the check were going to fail anyway even if pools actually exist.
Using jinja syntax to set `docker_exec_cmd` allows to handle the case
where you don't have monitors in your inventory.
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
If openstack_pools contains an application key it will be used to apply
this application pool type to a pool.
Closes: https://bugzilla.redhat.com/show_bug.cgi?id=1562220
Signed-off-by: Sébastien Han <seb@redhat.com>
As of ceph 12.2.5 the type of the parameter `type` is not a name anymore but
an id, therefore an `int` is expected otherwise it will fail with the
following error
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
The last mon creates the keys with a particular mode, while copying them
to the other mons (first and second) we must re-use the mode that was
set.
The same applies for the client node, the slurp preserves the initial
'item' so we can get the mode for the copy.
Signed-off-by: Sébastien Han <seb@redhat.com>
This key is created after the last mon is up so there is no need to try
to push it from the first mon. The initia mon container is not creating
the mgr key, ansible does. So this key will never exist.
The key will go into the fetch dir once the last mon is up, then when
the ceph-mgr plays it will try to get it from the fetch directory.
Signed-off-by: Sébastien Han <seb@redhat.com>
During the initial bootstrap of the first mon, the monmap file is
destroyed so it's not available and ansible will never find it.
Signed-off-by: Sébastien Han <seb@redhat.com>
Useful for softwares that do data collection/monitoring like collectd.
They can connect to the socket and then retrieve information.
Even though the sockets are exposed now, I'm keeping the docker exec to
check the socket, this will allow newer version of ceph-ansible to work
with older versions.
Closes: https://bugzilla.redhat.com/show_bug.cgi?id=1563280
Signed-off-by: Sébastien Han <seb@redhat.com>
We now have the ability to detect the uid/gid of the ceph user depending
on the distribution we are running on and so we are doing non-container
deployements.
Signed-off-by: Sébastien Han <seb@redhat.com>
We know bindmount with the :z option at the end of the -v command so
this will basically run the exact same command as we used to run. So to
speak:
chcon -Rt svirt_sandbox_file_t /var/lib/ceph
Signed-off-by: Sébastien Han <seb@redhat.com>
This fixes the case where the playbook died and never removed the
container. So now, once the container exits it will remove itself from
the container list.
Closes: https://bugzilla.redhat.com/show_bug.cgi?id=1568157
Signed-off-by: Sébastien Han <seb@redhat.com>
If the user has set copy_admin_key to true we assume he/she wants to
import the key in Ceph and not only create the key on the filesystem.
Signed-off-by: Sébastien Han <seb@redhat.com>
ceph-authtool does not support raw arguements so we have to quote caps
declaration like this allow 'bla bla' instead of allow bla bla
Closes: https://bugzilla.redhat.com/show_bug.cgi?id=1568157
Signed-off-by: Sébastien Han <seb@redhat.com>
This commit does a couple of things:
* use a common.yml file that contains things that can be played on both
container and non-container
* refactor the ability to copy the admin key to the nodes
Signed-off-by: Sébastien Han <seb@redhat.com>
Red Hat is now using tags[3,latest] for image rhceph/rhceph-3-rhel7.
Because of this, the ceph_uid conditional passes for Debian
when 'ceph_docker_image_tag: latest' on RH deployments.
I've added an additional task to check for rhceph image specifically,
and also updated the RH family task for ceph/daemon [centos|fedora]tags.
Signed-off-by: Randy J. Martinez <ramartin@redhat.com>
When installing rhcs on Debian systems the red hat repos must have the
highest priority so we avoid packages conflicts and install the rhcs
version.
Closes: https://bugzilla.redhat.com/show_bug.cgi?id=1565850
Signed-off-by: Sébastien Han <seb@redhat.com>
There is no need to check for a running cluster n*nodes time in
`ceph-defaults` so let's add a `run_once: true` to save some resources
and time.
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
Regardless if the partition is 'ceph' or something else, we don't want
to be as strick as checking for a particular partition.
If the drive has a partition, we just don't do anything.
This solves the case where the server reboots, disks get a different
/dev/sda (node) allocation. In this case, prior to restarting the server
/dev/sda was an OSD, but now it's /dev/sdb and the other way around.
In such scenario, we will try to prepare the OSD and create a new
partition, so let's not mess around with devices that have partitions.
Closes: https://bugzilla.redhat.com/show_bug.cgi?id=1498303
Signed-off-by: Sébastien Han <seb@redhat.com>
allow_multimds will be officially deprecated in Mimic, specify it
only for all versions of Ceph where it was declared stable. Going
forward, specify only max_mds.
Signed-off-by: Douglas Fuller <dfuller@redhat.com>
NFS-ganesha cannot start is the nfs-server service
is running. This commit stops nfs-server in case it
is running on a (debian, redhat, suse) node before
the nfs-ganesha service starts up
fixes: https://bugzilla.redhat.com/show_bug.cgi?id=1508506
Signed-off-by: Ali Maredia <amaredia@redhat.com>
Add a variable, ceph_nfs_disable_caching, that if set to true
disables ganesha's directory and attribute caching as much as
possible.
Also, disable caching done by ganesha, when 'nfs_file_gw'
variable is true, i.e., when Ganesha is used as CephFS's gateway.
This is the recommended Ganesha setting as libcephfs already caches
information. And doing so helps avoid cache incoherency issues
especially with clustered ganesha over CephFS.
Fixes: https://tracker.ceph.com/issues/23393
Signed-off-by: Ramana Raja <rraja@redhat.com>
If people keep on using the mon_cap, osd_cap etc the playbook will
translate this old syntax on the flight.
Signed-off-by: Sébastien Han <seb@redhat.com>
backward compatibility with `ceph_mon_docker_interface` and
`ceph_mon_docker_subnet` was not working since there wasn't lookup on
`monitor_interface` and `public_network`
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
Prior to this change, if a user had ceph-test-12.2.1 installed, and
upgraded to ceph v12.2.3 or newer, the RPM upgrade process would
fail.
The problem is that the ceph-test RPM did not depend on an exact version
of ceph-common until v12.2.3.
In Ceph v12.2.3, ceph-{osdomap,kvstore,monstore}-tool binaries moved
from ceph-test into ceph-base. When ceph-test is not yet up-to-date, Yum
encounters package conflicts between the older ceph-test and newer
ceph-base.
When all users have upgraded beyond Ceph < 12.2.3, this is no longer
relevant.
According to our recent change, we now use "CentOS" as a latest
container image. We need to reflect this on the ceph_uid.
Signed-off-by: Sébastien Han <seb@redhat.com>
Tripleo deployment failed when the monitors not manged
by tripleo itself with:
FAILED! => {"msg": "list object has no element 0"}
The failing play item was introduced by
f46217b69a .
fixes: https://bugzilla.redhat.com/show_bug.cgi?id=1552327
Signed-off-by: Attila Fazekas <afazekas@redhat.com>
because of `serial: 1`, it can be an issue when the playbook is being
run on client nodes.
Since the refact of `ceph-client` we skip the role `ceph-defaults` on
every node except the first client node, it means that the task is not
going to be played because of `run_once: true`.
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
This commit refacts this role so we don't have to pull container image
on client nodes just to create pools and keys.
Fixes: https://bugzilla.redhat.com/show_bug.cgi?id=1550977
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
This seems to be a leftover.
This commit removes an unnecessary 'set linux permissions' on
`/var/lib/ceph`
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
This check is alone in `ceph-docker-common` since a previous code
refactor.
Moving this check in `ceph-defaults` allows us to run `ceph-clients`
without having to run `ceph-docker-common` even in non-containerized
deployment.
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
Prior to this patch, the certificates where being generated on a single
node only (because of the run_once: true). Thus certificates were not
distributed on all the gateway nodes.
This would require a second ansible run to work. This patches fix the
creation and keys's distribution on all the nodes.
Closes: https://bugzilla.redhat.com/show_bug.cgi?id=1540845
Signed-off-by: Sébastien Han <seb@redhat.com>
This update will resolve error['cephfs' is undefined.] in multimds container deployments.
See: roles/ceph-mon/tasks/create_mds_filesystems.yml. The same last two tasks are present there, and actully need to happen in that role since "{{ cephfs }}" gets defined in
roles/ceph-mon/defaults/main.yml, and not roles/ceph-mds/defaults/main.yml.
Signed-off-by: Randy J. Martinez <ramartin@redhat.com>
Copy the admin key when configured nfs_file_gw (but not nfs_obj_gw). Also,
copy/setup RGW related directories only when configured as nfs_obj_gw.
Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
This variable is needed for containerized clusters and is required for
the ceph-docker-common role. Typically the is_atomic variable is set in
site-docker.yml.sample though so if ceph-docker-common is used outside
of that playbook it needs set in another way. Moving the creation of
the variable inside this role means playbooks don't need to worry
about setting it.
fixes: https://bugzilla.redhat.com/show_bug.cgi?id=1558252
Signed-off-by: Andrew Schoen <aschoen@redhat.com>
Since the approach to creating a ceph.conf file has changed, and now
no-longer relies on assembling config file fragments in /etc/ceph/ceph.d
we can avoid the conf_overrides rendering on the local host and skip out
the tasks related to that, instead using just the config_template task
to configure the file directly.
When creating pools, it's crucial to expose all the options available as
part of the pool creation command. As explained in:
http://docs.ceph.com/docs/jewel/rados/operations/pools/
Signed-off-by: Sébastien Han <seb@redhat.com>
If OSDs don't restart normally we now also dump info of the crush map,
crush rules, crush tree and pools.
If the monitors don't restart normally we also print the socket status
by calling mon_status and quorum_status.
Signed-off-by: Sébastien Han <seb@redhat.com>
The `pools` dict defined in `roles/ceph-client/defaults/main.yml`
shouldn't have `{{ ceph_conf_overrides.global.osd_pool_default_pg_num
}}` as default value for `pgs` keys.
For instance, if you want some pools to be created but without explicitely
specifying the pgs for these pools (it means you want to use the
`osd_pool_default_pg_num`), you will be obliged to define
`{{ ceph_conf_overrides.global.osd_pool_default_pg_num }}` anyway while you
wanted to use the current default value already defined in the cluster which is
retrieved early in the playbook and stored in the
`{{ osd_pool_default_pg_num }}` fact.
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
Running the last portion (insert new default and add new default crush
tasks) of crush_rules.yml only on the last monitor is
wrong since ceph CLI calls usually end up on the master having the
quorum, which is by default the one with the lower IP.
So if we run the command and end up on another mon the creation will
happen on the default crush rule because the particular mon hasn't been
updated.
To fix this we remove the |last on the include and use run_once: true on
certain tasks, then we let the final two tasks run on all the monitors.
Signed-off-by: Sébastien Han <seb@redhat.com>
On releases after jewel the option
'osd_pool_default_crush_replicated_ruleset' does not exist anymore, it's
called osd_pool_default_crush_rule.
Signed-off-by: Sébastien Han <seb@redhat.com>
This was causing a lot of pain with the handlers. Also the
implementation was not ideal since we were assembling files. Everything
can now be done with the ceph_crush module so let's remove that.
Signed-off-by: Sébastien Han <seb@redhat.com>
Instead of creating the CRUSH hierarchy with Ansible tasks using the
command module we now rely on the ceph_crush module.
Signed-off-by: Sébastien Han <seb@redhat.com>
One could want to add new crush rules while keeping his current default rule.
Fixed it so that it works with all rules defined as "default: false". If multiple rules are defined as default (should not be) then the last rule listed in "crush_rules" is taken as default.
As part of fcba2c801a these vars were
removed and no longer do anything:
radosgw_dns_name
radosgw_resolve_cname
This patch removes them from the group_vars files and defaults/main.yml
If we now set copy_admin_key while running a containerized scenario, the
ceph admin key will be copied on the node.
Signed-off-by: Sébastien Han <seb@redhat.com>
In case the admin wasn't copied over to the node this command would
fail. So it's safer to run it from a monitor directly.
Signed-off-by: Sébastien Han <seb@redhat.com>
That task is failing on containerized deployment because `ceph:ceph`
doesn't exist.
The idea here is to use the `{{ ceph_uid }}` to set the ownerships for
the admin keyring when containerized_deployment.
Closes: https://bugzilla.redhat.com/show_bug.cgi?id=1540578
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
The nfs-ganesha package has been fixed as part of this commit:
963b6681df
Once the package is rebuilt this should be good to merge.
This reverts commit e88af3c4cb.
Previously it was necessary to provide a value (eventually an
empty string) for the "rule_name" key for each item in
openstack_pools. This change makes that optional and defaults to
empty string when not given.
Using updatedb -e doesnt make a permanent change, but will updatedb
without the passed path.
To make this change more permanent we should update the
/etc/updatedb.conf file to include /var/lib/ceph.
Don't merge this.
Test to see if we copy over the nfs-ganesha-lock.service.debian8 file
properly, whether the Xenial CI job will work.
The upstream download.ceph.com nfs-ganesha package should be fixed for
xenial (which is in progress).
This fact is already set in site-docker.yml so there's no need to check
it again in ceph-docker-common
Signed-off-by: Paul Bourke <paul.bourke@oracle.com>
This patch fixes an issue where if hosts have different service lists,
it will prevent restarting changes on services that run later on.
For example, hostA in the mons and rgws group would initiate a config
change and restart of services on all mons and rgws hosts, even though
a separate hostB (which is only in the rgws group) has not had its
configuration changed yet. Additionally, when the second host has its
coniguration changed as part of the ceph-rgw role, it will not initiate
a restart since its inventory name != the first hosts.
To fix this we should run the restart once (using run_once: True)
as long as the host has called the handler. This will ensure that even
if only 1 host has called the handler it will initiate a restart on all
hosts that have called the handler.
Additionally, we add a var that is set when the handler runs, this will
ensure that only hosts that have called the handler get restarted.
Includes minor fix to remove unrequired "inventory_hostname in
play_hosts" when: clause. This is no longer required since the handlers
were changed. The host calling the handler will be in play_hosts
already.
When used along with delegate, run_once does not belong well. Thus,
using | last always brings the desired result.
Signed-off-by: Sébastien Han <seb@redhat.com>
We now look for any excisting containers, if any we compare their
running image with the latest pulled container image.
For OSDs, we iterate over the list of running OSDs, this handles the
case where the first OSD of the list has been updated (runs the new
image) and not the others.
Closes: https://bugzilla.redhat.com/show_bug.cgi?id=1526513
Signed-off-by: Sébastien Han <seb@redhat.com>
Multipath disks have partitions with a different format than what
ceph-ansible currently supports, this update makes ceph-ansible
aware of that format so multipath disks can be used as OSDs
Signed-off-by: Caleb Boylan <caleb.boylan@ormuco.com>
Since Luminous we need to set the application tag for each pool,
otherwise a CEPH_WARNING is generated when the pools are in use.
We should assign the OpenStack pools to their default which would be
"rbd". When updating to Luminous this would happen automatically to the
vms, images, backups and volumes pools, but for new deploys this is not
the case.
osd_scenario does not exist in the ceph-default role so if we try to
play ceph-default on an OSD node, the playbook will fail with undefined
variable.
Signed-off-by: Sébastien Han <seb@redhat.com>
This commit fixes a bug that occurs especially for dmcrypt scenarios.
There is an issue where the 'disk_list' container can't reach the ceph
cluster because it's not launched with `--net=host`.
If this container can't reach the cluster, it will hang on this step
(when trying to retrieve the dm-crypt key) :
```
+common_functions.sh:448: open_encrypted_part(): ceph --cluster abc12 --name \
client.osd-lockbox.9138767f-7445-49e0-baad-35e19adca8bb --keyring \
/var/lib/ceph/osd-lockbox/9138767f-7445-49e0-baad-35e19adca8bb/keyring \
config-key get dm-crypt/osd/9138767f-7445-49e0-baad-35e19adca8bb/luks
+common_functions.sh:452: open_encrypted_part(): base64 -d
+common_functions.sh:452: open_encrypted_part(): cryptsetup --key-file \
-luksOpen /dev/sdb1 9138767f-7445-49e0-baad-35e19adca8bb
```
It means the `ceph-run-osd.sh` script won't be able to start the
`osd_disk_activate` process in ceph-container because he won't have
filled the `$DOCKER_ENV` environment variable properly.
Adding `--net=host` to the 'disk_list' container fixes this issue.
Fixes: https://bugzilla.redhat.com/show_bug.cgi?id=1543284
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
While hostname -f will always return an hostname including its
domain part and -s without the domain part, the behavior when
no arguments are given can include or not include the domain part
depending on how the system is configured; the socket name might
not match the instance name then.
Was called too early, container was not yet started so the commands failed.
Moved the section after include docker/main.yml
Signed-off-by: Greg Charot <gcharot@redhat.com>
Use a nicer syntax for `local_action` tasks.
We used to have oneliner like this:
```
local_action: wait_for port=22 host={{ hostvars[inventory_hostname]['ansible_default_ipv4']['address'] }} state=started delay=10 timeout=500 }}
```
The usual syntax:
```
local_action:
module: wait_for
port: 22
host: "{{ hostvars[inventory_hostname]['ansible_default_ipv4']['address'] }}"
state: started
delay: 10
timeout: 500
```
is nicer and kind of way to keep consistency regarding the whole
playbook.
This also fix a potential issue about missing quotation :
```
Traceback (most recent call last):
File "/tmp/ansible_wQtWsi/ansible_module_command.py", line 213, in <module>
main()
File "/tmp/ansible_wQtWsi/ansible_module_command.py", line 185, in main
rc, out, err = module.run_command(args, executable=executable, use_unsafe_shell=shell, encoding=None, data=stdin)
File "/tmp/ansible_wQtWsi/ansible_modlib.zip/ansible/module_utils/basic.py", line 2710, in run_command
File "/usr/lib64/python2.7/shlex.py", line 279, in split
return list(lex) File "/usr/lib64/python2.7/shlex.py", line 269, in next
token = self.get_token()
File "/usr/lib64/python2.7/shlex.py", line 96, in get_token
raw = self.read_token()
File "/usr/lib64/python2.7/shlex.py", line 172, in read_token
raise ValueError, "No closing quotation"
ValueError: No closing quotation
```
writing `local_action: shell echo {{ fsid }} | tee {{ fetch_directory }}/ceph_cluster_uuid.conf`
can cause trouble because it's complaining with missing quotes, this fix solves this issue.
Fixes: https://bugzilla.redhat.com/show_bug.cgi?id=1510555
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
With two public networks configured - we found that with
"NETWORK_ADDR_1, NETWORK_ADDR_2" install process consistently became
broken, trying to find docker registry on second network, and not
finding mon container.
but without spaces
"NETWORK_ADDR_1,NETWORK_ADDR_2" install succeeds
so, containerized install is more peculiar with formatting of this line
Closes: https://bugzilla.redhat.com/show_bug.cgi?id=1534003
Signed-off-by: Sébastien Han <seb@redhat.com>
Description of problem: The 'get osd id' task goes through all the 10 times (and its respective timeouts) to make sure that the number of OSDs in the osd directory match the number of devices.
This happens always, regardless if the setup and deployment is correct.
Version-Release number of selected component (if applicable): Surely the latest. But any ceph-ansible version that contains ceph-volume support is affected.
How reproducible: 100%
Steps to Reproduce:
1. Use ceph-volume (LVM) to deploy OSDs
2. Avoid using anything in the 'devices' section
3. Deploy the cluster
Actual results:
TASK [ceph-osd : get osd id _uses_shell=True, _raw_params=ls /var/lib/ceph/osd/ | sed 's/.*-//'] **********************************************************************************************************************************************
task path: /Users/alfredo/python/upstream/ceph/src/ceph-volume/ceph_volume/tests/functional/lvm/.tox/xenial-filestore-dmcrypt/tmp/ceph-ansible/roles/ceph-osd/tasks/start_osds.yml:6
FAILED - RETRYING: get osd id (10 retries left).
FAILED - RETRYING: get osd id (9 retries left).
FAILED - RETRYING: get osd id (8 retries left).
FAILED - RETRYING: get osd id (7 retries left).
FAILED - RETRYING: get osd id (6 retries left).
FAILED - RETRYING: get osd id (5 retries left).
FAILED - RETRYING: get osd id (4 retries left).
FAILED - RETRYING: get osd id (3 retries left).
FAILED - RETRYING: get osd id (2 retries left).
FAILED - RETRYING: get osd id (1 retries left).
ok: [osd0] => {
"attempts": 10,
"changed": false,
"cmd": "ls /var/lib/ceph/osd/ | sed 's/.*-//'",
"delta": "0:00:00.002717",
"end": "2018-01-21 18:10:31.237933",
"failed": true,
"failed_when_result": false,
"rc": 0,
"start": "2018-01-21 18:10:31.235216"
}
STDOUT:
0
1
2
Expected results:
There aren't any (or just a few) timeouts while the OSDs are found
Additional info:
This is happening because the check is mapping the number of "devices" defined for ceph-disk (in this case it would be 0) to match the number of OSDs found.
Basically this line:
until: osd_id.stdout_lines|length == devices|unique|length
Means in this 2 OSD case it is trying to ensure the following incorrect condition:
until: 2 == 0
Closes: https://bugzilla.redhat.com/show_bug.cgi?id=1537103
This should default to False. The default for Keystone is not to use PKI
keys, additionally, anybody using this setting had to have been manually
setting it before.
Fixes: #2111
This allows us to use host-specific variables in ceph_conf_overrides variable. For example, this fixes usage of such variables (e.g. 'nss db path' having {{ ansible_hostname }} inside) in ceph_conf_overrides for rados gateway configuration (see profiles/rgw-keystone-v3) - issue #2157.
Signed-off-by: Eduard Egorov <eduard.egorov@icl-services.com>
Sometime the playbook gets stuck because even with `--connect-timeout=`
option, the connexion to the existing ceph cluster never timeout.
As a workaround, using `timeout` command provided by coreutils will
actually timeout if we can't connect to the cluster.
Closes: https://bugzilla.redhat.com/show_bug.cgi?id=1537003
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
This is to keep backward compatibility with stable-2.2 and satisfy the
check "verify dedicated devices have been provided" in
`check_mandatory_vars.yml`. This check is looking for
`dedicated_devices` so we need to default it's value to
`raw_journal_devices` when `raw_multi_journal` is set to `True`.
Fixes: https://bugzilla.redhat.com/show_bug.cgi?id=1536098
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
Some systems that were deployed with old tools can leave units named
"ceph-radosgw@radosgw.gateway.service". As a consequence, they will
prevent the new unit to start.
Closes: https://bugzilla.redhat.com/show_bug.cgi?id=1509584
Signed-off-by: Sébastien Han <seb@redhat.com>
Currently, we can define crush location for each host but only crush roots and crush rules are created. This commit automates other routines for a complete solution:
1) Creates rack type crush buckets defined in {{ ceph_crush_rack }} of each osd host. If it's not defined by user then a rack named 'default_rack_{{ ceph_crush_root }}' would be added and used in next steps.
2) Move rack type crush buckets defined in {{ ceph_crush_rack }} into crush roots defined in {{ ceph_crush_root }} of each osd host.
3) Move hosts defined in {{ ceph_crush_rack }} into crush roots defined in {{ ceph_crush_root }} of each osd host.
Signed-off-by: Eduard Egorov <eduard.egorov@icl-services.com>
On a non-collocated scenario, if a drive is faulty we can't really
remove it from the list of 'devices' without messing up or having to
re-arrange the order of the 'dedicated_devices'. We want to keep this
device list ordered. This will prevent the activation failing on a
device that we know is failing but we can't remove it yet to not mess up
the dedicated_devices mapping with devices.
Signed-off-by: Sébastien Han <seb@redhat.com>
Having handlers in both ceph-defaults and ceph-docker-common roles can make the
playbook restarting two times services. Handlers can be triggered first
time because of a change in ceph.conf and a second time because a new
image has been pulled.
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
This wasn't any good choice to implement this.
We had several options and none of them were ideal since handlers can
not be triggered cross-roles.
We could have achieved that by doing:
* option 1 was to add a dependancy in the meta of the ceph-docker-common
role. We had that long ago and we decided to stop so everything is
managed via site.yml
* option 2 was to import files from another role. This is messy and we
don't that anywhere in the current code base. We will continue to do so.
There is option 3 where we pull the image from the ceph-config role.
This is not suitable as well since the docker command won't be available
unless you run Atomic distro. This would also mean that you're trying to
pull twice. First time in ceph-config, second time in ceph-docker-common
The only option I came up with was to duplicate a bit of the ceph-config
handlers code.
Closes: https://bugzilla.redhat.com/show_bug.cgi?id=1526513
Signed-off-by: Sébastien Han <seb@redhat.com>
When containerized deployment, `docker_exec_cmd` is not set before the
task which try to retrieve the current fsid is played, it means it
considers there is no existing fsid and try to generate a new one.
Typical error:
```
ok: [mon0 -> mon0] => {
"changed": false,
"cmd": [
"ceph",
"--connect-timeout",
"3",
"--cluster",
"test",
"fsid"
],
"delta": "0:00:00.179909",
"end": "2018-01-09 10:36:58.759846",
"failed": false,
"failed_when_result": false,
"rc": 1,
"start": "2018-01-09 10:36:58.579937"
}
STDERR:
Error initializing cluster client: Error('error calling conf_read_file: errno EINVAL',)
```
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
Previously we were using ceph_conf_overrides however this doesn't play
nice for softwares like TripleO that uses ceph_conf_overrides inside its
own code. For now, and since this is the only occurence of this, we can
ensure no logs through the ceph conf template.
Closes: https://bugzilla.redhat.com/show_bug.cgi?id=1532619
Signed-off-by: Sébastien Han <seb@redhat.com>
There is no reasons why we can't use crush rules when deploying
containers. So moving the inlcude in the main.yml so it can be called.
Signed-off-by: Sébastien Han <seb@redhat.com>
ceph-create-keys is idempotent so it's not an issue to run it each time
we play ansible. This also fix issues where the 'creates' arg skips the
task and no keys get generated on newer version, e.g during an upgrade.
Closes: https://github.com/ceph/ceph-ansible/issues/2228
Signed-off-by: Sébastien Han <seb@redhat.com>
When upgrading from OSP11 to OSP12 container, ceph-ansible attempts to
disable the RGW service provided by the overcloud image. The task
attempts to stop/disable ceph-rgw@{{ ansible-hostname }} and
ceph-radosgw@{{ ansible-hostname }}.service. The actual service name is
ceph-radosgw@radosgw.$name
Closes: https://bugzilla.redhat.com/show_bug.cgi?id=1525209
Signed-off-by: Sébastien Han <seb@redhat.com>
the gpt label creation doesn't work even with parted module.
This commit fixes the gpt label creation by using parted command
instead.
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
We have a scenario when we switch from non-container to containers. This
means we don't know anything about the ceph partitions associated to an
OSD. Normally in a containerized context we have files containing the
preparation sequence. From these files we can get the capabilities of
each OSD. As a last resort we use a ceph-disk call inside a dummy bash
container to discover the ceph journal on the current osd.
Closes: https://bugzilla.redhat.com/show_bug.cgi?id=1525612
Signed-off-by: Sébastien Han <seb@redhat.com>
This resolves the following error:
E: There were unauthenticated packages and -y was used without
--allow-unauthenticated
Signed-off-by: Sébastien Han <seb@redhat.com>
The name docker_version is very generic and is also used by other
roles. As a result, there may be name conflicts. To avoid this a
ceph_ prefix should be used for this fact. Since it is an internal
fact renaming is not a problem.
making `osd_pool_default_pg_num` mandatory is a bit agressive and is
unrelated when you just want to create users keyrings.
Closes: #2241
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
the entrypoint to generate users keyring is `ceph-authtool`, therefore,
it can expand the `$(ceph-authtool --gen-print-key)` inside the
container. Users must generate a keyring themselves.
This commit also adds a check to ensure keyring are properly filled when
`user_config: true`.
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
Add the variables ceph_osd_docker_cpuset_cpus and
ceph_osd_docker_cpuset_mems, so that a user may specify
the CPUs and memory nodes of NUMA systems on which OSD
containers are run.
Provides a example in osds.yaml.sample to guide user
based on sample `lscpu` output since cpuset-mems refers
to the memory by NUMA node only while cpuset-cpus can
refer to individual vCPUs within a NUMA node.
If a deployer uses an interface name with a dash/hyphen in it, such
as 'br-storage' for the monitor_interface group_var, the ceph.conf.j2
template fails to find the right facts. It looks for
'ansible_br-storage' but only 'ansible_br_storage' exists.
This patch converts the interface name to underscores when the
template does the fact lookup.
The CI complains because of `ceph_uid` fact which doesn't exist since
the docker image tag used in the CI doesn't match with this condition.
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
This is particularly useful in CI environments where you dont have
the option of adding extra devices or volumes to the host. It is also
a simple change to support loopback devices
In case where docker CLI is available but docker is not running, we
don't want to trigger the restart of the daemons.
Closes: https://bugzilla.redhat.com/show_bug.cgi?id=1510555
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
Since some daemons now install their own packages the task checking the
ceph version fails on Debian systems. So the 'ceph-common' package must
be installed on all the machines.
Signed-off-by: Sébastien Han <seb@redhat.com>
This task was originally needed to fix a docker installation issue
(see: #1030). This has been fixed, therefore it can be removed.
Fixes: #2199
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
A recent change [1] required that the openstack_keys
param always containe an acls list. However, it's
possible it might not contain that list. Thus, this
param sets a default for that list to be empty if it
is not in the structure as defined by the user.
[1] d65cbaa539
We were activating dmcrypt devices with the wrong command. Basically the
first task execute the wrong activate command. The task fails but
continues because of the 'failed_when: false'. Then the right activation
sequence is being done by the next task.
Signed-off-by: Sébastien Han <seb@redhat.com>
when `ceph-rbd-mirror.target` is not enabled, the service won't start
after a reboot because there is a dependency between these two units.
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
If ceph-ansible deploys a Ceph cluster with "openstack_config: true"
and sets the openstack_keys map to have certain ACLs or permissions,
the requested ACLs or permissions are only set on one of the monitor
nodes [2] when they should be set on all of them.
This patch solves [3] the above issue by having the chmod and setfacl
tasks iterate the list of mon nodes (including the mon node that the
task was delegated to) to apply the chmod of setfacl to the keys in
openstack_keys.
[1]
```
openstack_keys:
- { name: client.openstack, key: "$(ceph-authtool --gen-print-key)", mon_cap: "allow r", osd_cap: "allow class-read object_prefix rbd_children, allow rwx pool=images, allow rwx pool=vms, allow rwx pool=volumes, allow rwx pool=backups", mode: "0600", acls: ["u:nova:r--", "u:cinder:r--", "u:glance:r--", "u:gnocchi:r--"] }
```
[2]
```
$ ansible mons -m shell -b -a "ls -l /etc/ceph/ceph.client.openstack.keyring ; getfacl /etc/ceph/ceph.client.openstack.keyring"
192.168.1.26 | SUCCESS | rc=0 >>
-rw-r-----+ 1 root root 253 Nov 3 20:30 /etc/ceph/ceph.client.openstack.keyring
user::rw-
user:glance:r--
user:nova:r--
user:cinder:r--
user:gnocchi:r--
group::---
mask::r--
other::---getfacl: Removing leading '/' from absolute path names
192.168.1.29 | SUCCESS | rc=0 >>
-rw-r--r--. 1 root root 253 Nov 3 20:30 /etc/ceph/ceph.client.openstack.keyring
user::rw-
group::r--
other::r--getfacl: Removing leading '/' from absolute path names
192.168.1.23 | SUCCESS | rc=0 >>
-rw-r--r--. 1 root root 253 Nov 3 20:30 /etc/ceph/ceph.client.openstack.keyring
user::rw-
group::r--
other::r--getfacl: Removing leading '/' from absolute path names
$
```
[3]
```
(undercloud) [stack@hci-director ceph-ansible]$ ansible mons -m shell -b -a "ls -l /etc/ceph/ceph.client.openstack.keyring ; getfacl /etc/ceph/ceph.client.openstack.keyring"
192.168.1.25 | SUCCESS | rc=0 >>
-rw-r-----+ 1 root root 253 Nov 14 01:12 /etc/ceph/ceph.client.openstack.keyring
user::rw-
user:glance:r--
user:nova:r--
user:cinder:r--
user:gnocchi:r--
group::---
mask::r--
other::---getfacl: Removing leading '/' from absolute path names
192.168.1.29 | SUCCESS | rc=0 >>
-rw-r-----+ 1 root root 253 Nov 14 01:12 /etc/ceph/ceph.client.openstack.keyring
user::rw-
user:glance:r--
user:nova:r--
user:cinder:r--
user:gnocchi:r--
group::---
mask::r--
other::---getfacl: Removing leading '/' from absolute path names
192.168.1.27 | SUCCESS | rc=0 >>
-rw-r-----+ 1 root root 253 Nov 14 01:12 /etc/ceph/ceph.client.openstack.keyring
user::rw-
user:glance:r--
user:nova:r--
user:cinder:r--
user:gnocchi:r--
group::---
mask::r--
other::---getfacl: Removing leading '/' from absolute path names
(undercloud) [stack@hci-director ceph-ansible]$
```
Add support for the openSUSE distributions. The required packages
are available either in the distribution repositories or in the
OBS one.
Signed-off-by: Markos Chandras <mchandras@suse.de>
When we consume the distribution packages, we don't have the choise on
which version to install, so we shouldn't require that variable to be
set. Distributions normally provide only one version of Ceph in the
official repositories so we get whatever they provide.
Signed-off-by: Markos Chandras <mchandras@suse.de>
openSUSE Leap 42.3 provides support for Ceph Luminous in both the
distribution package and the latest available version in the OBS
repository so add these as the only available installation methods for
openSUSE.
Signed-off-by: Markos Chandras <mchandras@suse.de>
Like 80d32dec, the path to the fact is not correct.
In any case, we will retrieve the IP address in hostvars, the variable
is the way we get the interface name according where it has been set
(eg.: inventory host file vs. group_vars/)
Closes: https://bugzilla.redhat.com/show_bug.cgi?id=1510906
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
there is no need to have a condition on this task, this test should be
always run since the result will be interpreted later.
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
This will prevent ceph-ansible from using a loop device while it
shouldn't in auto_discovery mode.
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
The path to the fact is not correct.
In any case, we will retrieve the IP address in hostvars, the variable
is the way we get the interface name according where it has been set
(eg.: inventory host file vs. group_vars/)
Closes: https://bugzilla.redhat.com/show_bug.cgi?id=1510906
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
Use `devices` variable instead of `ansible_devices`, otherwise it means
we are not using the devices which have been 'auto discovered'
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
The current code will also return lvm devices such as /dev/dm-2, this
kind of device type is not supported by ceph-disk at the moment. Now we
just ignore them.
Signed-off-by: Sébastien Han <seb@redhat.com>
- One can not run scripts directly in place, that mounted with `noexec`
option. But one can run scripts as arguments for `bash/sh`.
Signed-off-by: Arano-kai <captcha.is.evil@gmail.com>
During the initial implementation of this 'old' thing we were falling
into this issue without noticing
https://github.com/moby/moby/issues/30341 and where blindly using --rm,
now this is fixed the prepare container disappears and thus activation
fail.
I'm fixing this for old jewel images.
Also this fixes the machine reboot case where the docker logs are
purgend. In the old scenario, we now store the log locally in the same
directory as the ceph-osd-run.sh script.
Signed-off-by: Sébastien Han <seb@redhat.com>
Setting monitor_interface in group_vars/all.yml makes the
hostvars[host]['monitor_interface'] non-existing.
Closes: https://bugzilla.redhat.com/show_bug.cgi?id=1507922
Signed-off-by: Sébastien Han <seb@redhat.com>
Only chmod or setfacl the requested keyring(s) in the
opentack_keys data structure when the mode or acls keys
of that data structure exist.
User may specify four permission combinations for the
keyring file(s): 1. only set ACL, 2. only set mode,
3. set neither mode nor ACL, 4. set mode and then ACL.
Fixes: #2092
Use "ceph_tcmalloc_max_total_thread_cache" to set the
TCMALLOC_MAX_TOTAL_THREAD_CACHE_BYTES value inside /etc/default/ceph for
Debian installs, or /etc/sysconfig/ceph for Red Hat/CentOS installs.
By default this is set to 0, so the default package value will be used,
if specified this value will be changed to match the variable, and ceph
osd services will be restarted.
There was a huge resync from luminous to jewel in ceph-docker:
https://github.com/ceph/ceph-docker/pull/797
This change brought a new handy function to discover partitions tight to
an OSD. This function doesn't exist in the old image so the
ceph-osd-run.sh script breaks when trying to deploy Jewel OSD with that
old Jewel image version.
Signed-off-by: Sébastien Han <seb@redhat.com>
stable-3.0 brought numerous changes in ceph-ansible variables, this PR
aims to maintain backward compatibility for someone running stable-2.2
upgrading to stable-3.0 but keeps its groups_vars untouched.
We will then determine the right options to make sure the upgrade works
but we are expecting that new variables should be used.
We will drop this in a near future, maybe 3.1 or 3.2.
Signed-off-by: Sébastien Han <seb@redhat.com>
This feature isn't available before luminous, therefore, we need to play
them only on luminous and after otherwise the playbook will fail.
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
(cherry picked from commit 3f3d4b9c727d06154c422d445fc2a245aceaed89)
3a58757 introduced an issue for Jewel deployments, since this role is
skipped, `enabled_ceph_mgr_modules.stdout` doesn't exist, therefore, it
ends up with an attribute error.
Uses `.get()` to retrieve `stdout` with a default value so it won't fail
if this attribute doesn't exist (jewel).
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
In Jewel, we don't use bootstrap-rbd keyring for rbd-mirror nodes, it
results with a socket path/name different according to which ceph
release you are deploying.
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
This commit add new osd scenarios, it aims to simplify the CI setup and
brings a better coverage on the OSD scenarios.
We decided to differentiate between filestore and bluestore, thinking
ahead when filestore won't be supported anymore.
So we now have two classes of tests:
* Filestore
* Bluestore
In each of those classes we have container and non-container.
Then for each we test the following:
* collocated
* collocated dmcrypt
* non-collocated
* non-collocated dmcrypt
* auto discovery collocated
* auto discovery collocated dmcrypt
This gives us a nice coverage and also reduces the footprint on the CI.
We are now up to 4 scenarios, each containing 6 OSD VMs.
Signed-off-by: Sébastien Han <seb@redhat.com>
When doing collocation the condition "inventory_hostname in play_hosts"
is breaking the restart workflow.
Signed-off-by: Sébastien Han <seb@redhat.com>
- Add upgrade support for rbd mirror and nfs daemons.
- Only works with systemd (remove sysvinit and upstart occurence)
- A bit of cleanup
Signed-off-by: Sébastien Han <seb@redhat.com>
This will solve the following issue when starting docker containers on ubuntu:
invalid argument "1\u00a0" for --cpus=1 : failed to parse 1 as a rational number
Closes-bug: #2056
1. add the variables to docker_collocation
2. trigger the check when a MDS is part of the inventory file, not when
we run on an MDS...
Signed-off-by: Sébastien Han <seb@redhat.com>
The container deployment is serialized, adding this task as a best
effort. If docker is already present we pull the image otherwise we wait
for the role to play.
Signed-off-by: Sébastien Han <seb@redhat.com>
on jewel `ceph-rbd-mirror.target` isn't enabled, therefore, if the node
is rebooted, the service doesn't get started.
from ceph-rbd-mirror unit file:
```
[Install]
WantedBy=ceph-rbd-mirror.target
```
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
This patch simplifies the checks and installation tasks for NTP.
Debian and Red Hat had a check for NTP's presence but would then
install NTP right afterwards anyways. In addition, there were
tasks for atomic that weren't used anywhere else in the role.
This patch also uses a dynamic include to reduce delays from
skipped tasks.
This patch changes the `when:` keys so that they have no jinja2
delimiters. This avoids Ansible warnings which could turn into
errors in a future Ansible release.
We now have a variable called ceph_pools that is mandatory when
deploying a MDS.
It's a dictionnary that contains a pool name and a PG count. PG count is
mandatory and must be set, the playbook will fail otherwise.
Closes: https://github.com/ceph/ceph-ansible/issues/2017
Signed-off-by: Sébastien Han <seb@redhat.com>
Ansible throws warnings when using yum/dnf/rpm with the command
module:
[WARNING]: Consider using yum module rather than running yum
This patch adds the `warn: no` argument to suppress the warnings
in the Ansible output.
The `always_run` key is deprecated and being removed in Ansible 2.4.
Using it causes a warning to be displayed:
[DEPRECATION WARNING]: always_run is deprecated.
This patch changes all instances of `always_run` to use the `always`
tag, which causes the task to run each time the playbook runs.
Modern versions of Ansible can handle a list of packages passed
directly to the package modules. This patch optimizes the package
install process by passing the list of packages directly to the
module.
This is causing unknown issues when trying to start a dmcrypt container.
Basically the container is stuck at mount opening the LUKS device. This
is still unknown why this is causing trouble but we need to move
forward. Also, this doesn't seem to help in any ways to fix the race
condition we've seen.
Here is the log for dmcrypt:
cryptsetup 1.7.4 processing "cryptsetup --debug --verbose --key-file
key luksClose fbf8887d-8694-46ca-b9ff-be79a668e2a9"
Running command close.
Locking memory.
Installing SIGINT/SIGTERM handler.
Unblocking interruption on signal.
Allocating crypt device context by device
fbf8887d-8694-46ca-b9ff-be79a668e2a9.
Initialising device-mapper backend library.
dm version [ opencount flush ] [16384] (*1)
dm versions [ opencount flush ] [16384] (*1)
Detected dm-crypt version 1.14.1, dm-ioctl version 4.35.0.
Device-mapper backend running with UDEV support enabled.
dm status fbf8887d-8694-46ca-b9ff-be79a668e2a9 [ opencount flush ]
[16384] (*1)
Releasing device-mapper backend.
Trying to open and read device /dev/sdc1 with direct-io.
Allocating crypt device /dev/sdc1 context.
Trying to open and read device /dev/sdc1 with direct-io.
Initialising device-mapper backend library.
dm table fbf8887d-8694-46ca-b9ff-be79a668e2a9 [ opencount flush
securedata ] [16384] (*1)
Trying to open and read device /dev/sdc1 with direct-io.
Crypto backend (gcrypt 1.5.3) initialized in cryptsetup library
version 1.7.4.
Detected kernel Linux 3.10.0-693.el7.x86_64 x86_64.
Reading LUKS header of size 1024 from device /dev/sdc1
Key length 32, device size 1943016847 sectors, header size 2050
sectors.
Deactivating volume fbf8887d-8694-46ca-b9ff-be79a668e2a9.
dm status fbf8887d-8694-46ca-b9ff-be79a668e2a9 [ opencount flush ]
[16384] (*1)
Udev cookie 0xd4d14e4 (semid 32769) created
Udev cookie 0xd4d14e4 (semid 32769) incremented to 1
Udev cookie 0xd4d14e4 (semid 32769) incremented to 2
Udev cookie 0xd4d14e4 (semid 32769) assigned to REMOVE task(2) with
flags (0x0)
dm remove fbf8887d-8694-46ca-b9ff-be79a668e2a9 [ opencount flush
retryremove ] [16384] (*1)
fbf8887d-8694-46ca-b9ff-be79a668e2a9: Stacking NODE_DEL [verify_udev]
Udev cookie 0xd4d14e4 (semid 32769) decremented to 1
Udev cookie 0xd4d14e4 (semid 32769) waiting for zero
Signed-off-by: Sébastien Han <seb@redhat.com>
in addition to c4dcdaa20 this commit adds the missing condition on
install tasks for debian_rhcs deployment. Without them, these tasks are
played on any kind of deployment.
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
* DBus on host should include ganesha service file
* to allow ganesha container to respond on DBus it needs to run
in --privileged mode (ganesha folks contacted to look at this)
* ceph_nfs_include_exports_dir variable replaced with more general
ceph_nfs_dynamic_exports
This is something that has nothing to do in `ceph-common`, this
is too specific to `ceph-iscsi-gw` role.
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
This is something that has nothing to do in `ceph-common`, this
is too specific to `ceph-nfs` role.
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
Make role `ceph-mgr` handling itself the installation of `ceph-mgr`
package because it's complicated to manage it regarding we are going to
install `jewel vs. luminous`
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
During the initial play, the docker command doesn't not exist and then
there is no stdout_lines to the command. So get allows us to fix this by
declaring an array if the command fails.
Signed-off-by: Sébastien Han <seb@redhat.com>
Use an intermediate variable to build the final `dedicated_devices` list
to avoid duplicate entry in that array. (We need a 1:1 relation between
`dedicated_devices` and `devices` since we are using a `with_together`
later.
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
With jewel, `bootstrap_rbd_keyring` is not set because of this condition:
```
when:
- ceph_release_num.{{ ceph_release }} >= ceph_release_num.luminous
```
Therefore, the task `try to fetch ceph config and keys` will fail.
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
* Change version from 2 to 3.
* use ceph_rhcs_cdn_debian_repo_version to use other repositories along
* with ceph_rhcs_cdn_debian_repo
Signed-off-by: Sébastien Han <seb@redhat.com>
`ceph_release` isn't available at this step of the playbook because it
is set later based on the installed binaries.
Closes: https://bugzilla.redhat.com/show_bug.cgi?id=1486062
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
The value of doing this is fairly low compare to the added value.
So we remove these tasks, if rbd pool on Jewel doesn't have the right PG
value you can always increase it.
Signed-off-by: Sébastien Han <seb@redhat.com>
All keyring are getting copied to all nodes.
This commit fixes a leftover from a previous code refactor.
Closes: https://bugzilla.redhat.com/show_bug.cgi?id=1498583
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
iscsi-gw needs a 'rbd' pool to configure iscsi target.
Note: I could have used the facts already set in `ceph-mon` but I voluntarily
didn't do it to not create a dependancy between these two roles.
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
This commit refacts the code regarding all `set_osd_pool_default_*`
related tasks by avoiding usage of useless `set_fact` to determine
whether a key is present in `ceph_conf_overrides`.
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
The rbd pool is the default pool that gets created during ceph cluster
initializaiton. If we act on the rbd related operations too early, the
rbd pool does not exist yet. Move the call to perform rbd operations
to a later stage after other pools have been created.
The rbd_pool.yml playbook has all the operations related to the rbd pool.
Replace the always_run (deprecated) directive with check_mode.
Most of the ceph related tasks only need to run once. The run_once directive
executes the task on the first host.
The ceph sub-command to delete a pool is delete (not rm).
The changes submitted here were tested with this ceph version.
ceph version 0.94.9-9.el7cp (b83334e01379f267fb2f9ce729d74a0a8fa1e92c)
This upload includes these changes:
- Use the fail module (instead of assert).
- From luminous release, the rbd pool is no longer created by default.
Delete the code to create the rbd pool for luminous release
- Conform the .yml files to use the suggested syntax.
The commands are executed on the mcp nodes and I think shell ansible module
is the right one to use. The command module is used to execute commands on
remote nodes. I can make the change to use command module if that is
prefrerred.
This is to ensure `docker_exec_cmd` fact is set with the correct value
in case of daemons collocation
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
Also we now play ceph-config to have everything being generated for new
daemons bootstrap during upgrade.
Closes: https://bugzilla.redhat.com/show_bug.cgi?id=1497959
Signed-off-by: Sébastien Han <seb@redhat.com>
generate_crt|bool|default(false) won't apply the default value, this
generate_crt|default(false)|bool will
Signed-off-by: Sébastien Han <seb@redhat.com>
The task which sets `ceph_current_fsid` in `facts.yml` in case of containerized
deployment, will definitely fail because `docker_exec_cmd` is not set
yet.
This commits simply makes `facts.yml` played after `check_socket.yml` so
`docker_exec_cmd` is set properly.
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
This commits refacts the role ceph-mds
The goal here is to create cephfs in `ceph-mon` for both containerized
and non-containerized cases so we don't need the admin keyring on mds
nodes anymore.
Closes: https://bugzilla.redhat.com/show_bug.cgi?id=1488999
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
Using systemd module allows us to do in one task what we did in three
tasks:
- enable unit file,
- issue a `daemon-reload`,
- start the service
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
Running the socket check on all the hosts will override the default
value of docker_exec_cmd, leaving it with the last value (currently
rbd-mirror), as a result the subsequent docker_exec_cmd usage for the
:x
Signed-off-by: Sébastien Han <seb@redhat.com>
There is a bug in the rbd mirror unit file, the upstream fix is here:
https://github.com/ceph/ceph/pull/17969.
This should be reverted once the patch is merged and backport is done.
Signed-off-by: Sébastien Han <seb@redhat.com>
This fixes the error :
```
The conditional check 'sestatus.stdout != 'Disabled'' failed.
```
that occurs when running on non rhel based system since the
`sestatus` fact is registered only on rhel based distribution.
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
Specify the timeout flag to ceph-create-keys, which causes it to time out
if a monitor quorum isn't achieved. This overrides the default timeout
of 10 minutes, causing ceph-ansible to fail faster in the event of cluster
network issues.
Signed-off-by: Douglas Fuller <dfuller@redhat.com>
We have seen issues with leftover socker. So now, if a socket is found
we also check if it's accessed by a process. If so, we can run the
handler, if not we remove it and continue the playbook.
Signed-off-by: Sébastien Han <seb@redhat.com>
Co-Authored-by: Guillaume Abrioux <gabrioux@redhat.com>
It's sad but we can not rely on the prepare container anymore since the
log are flushed after reboot. So inpecting the container does not return
anything.
Now, instead we use a ephemeral container to look up for the
journal/block.db/block.wal (depending if filestore or bluestore) and
build the activate command accordingly.
Signed-off-by: Sébastien Han <seb@redhat.com>
need to use `hostvars[host]['XXX']` to retrieve the monitor
interface and/or radosgw interface.
Fixes: https://bugzilla.redhat.com/show_bug.cgi?id=1493920
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
- move the file fetch/push to the existing task
- rename the include
- generate the ganesha template from ansible
- re-arrange role structure
- re-use tasks for non-container and container
- configure keys for non-container and container
- fix rgw container key collection;
Signed-off-by: Sébastien Han <seb@redhat.com>
the rbd key was not pushed on rbd nodes because its keyring path was not
added in `ceph_config_keys`.
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
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>
- 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>
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>
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.
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.
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>
The variable "statleftover" was removed by commit
a60c74f61e
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>
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>
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>
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>
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>
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>
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>
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
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>
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>
This patch adds passing the RGW_CIVETWEB_IP to the docker
container. This IP defaults to the value of radosgw_civetweb_bind_ip.
radosgw_civetweb_bind_ip default to ipv4.default
Without this value, the RGW containter will bind to 0.0.0.0
Remove the old check prior systemd.
We only support systemd so there is no need to run a condition on
systemd. The playbook will fail if systemd is not present.
Signed-off-by: Sébastien Han <seb@redhat.com>
The installation process is now described as follow:
* you still have to choose a 'ceph_origin' installation method. The
origin can be a 'repository' (add a new repository), distro (it will use
the packages provided by the native repo source of your distribution),
local (only available on redhat system, it installs locally built
packages). This option is not well tested, so use it carefully
* if ceph_origin == 'repository' you will have to decide what kind of
repository you want to enable:
- community: corresponds to the stable upstream/community version
- enterprise: corresponds to the stable enterprise/downstream version
(basically you are a red hat customer)
- dev: it will install ceph from packages built out of the github
development branches
Signed-off-by: Sébastien Han <seb@redhat.com>
Co-Authored-by: Guillaume Abrioux <gabrioux@redhat.com>
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
The list can not be evaluated properly if it containers '[]', which is
the case when using the filter "default([])". To fix this, we have to
properly merge the lists.
This is fixing the issue: "list object has no element 1"
Signed-off-by: Sébastien Han <seb@redhat.com>
By detecting the ceph version running in the container we can easily
apply conditions like:
ceph_release_num.{{ ceph_release }} >= ceph_release_num.luminous
We do that already, in ceph-docker-common/tasks/fetch_configs.yml.
This fixes the error:
TASK [ceph-docker-common : register rbd bootstrap key]
******************************************************
fatal: [magna005]: FAILED! => {"failed": true, "msg": "The conditional
check 'ceph_release_num.{{ ceph_release }} >= ceph_release_num.luminous'
failed. The error was: error while evaluating conditional
(ceph_release_num.{{ ceph_release }} >= ceph_release_num.luminous):
'dict object' has no attribute 'dummy'\n\nThe error appears to have been
in
'/home/ubuntu/ceph-ansible/roles/ceph-docker-common/tasks/fetch_configs.yml':
line 2, column 3, but may\nbe elsewhere in the file depending on the
exact syntax problem.\n\nThe offending line appears to be:\n\n---\n-
name: register rbd bootstrap key\n ^ here\n"}
Closes: https://bugzilla.redhat.com/show_bug.cgi?id=1486062
Signed-off-by: Sébastien Han <seb@redhat.com>
Logging inside the container is not useful since it writes to the
overlayfs partition, resulting in potential performance degradation on
the container.
If you need to check the logs, just look at journald.
Signed-off-by: Sébastien Han <seb@redhat.com>
with_items is evaluated before the when condition so if the task that
registers the 'results' is skipped the task will fail with:
{"failed": true, "msg": "'dict object' has no attribute 'results'"}
Defaulting to an empty array fixes the issue.
Reverts: abdd66619e
Closes: https://bugzilla.redhat.com/show_bug.cgi?id=1482061
Signed-off-by: Sébastien Han <seb@redhat.com>
The script can fail to get the osd id because the osds are activated by
udev and it can take a while for them to activate. This commit fixes
that by trying to get all the osds per node in a loop.
This commit also makes the osd services enabled so that they are
available after reboot.
Signed-off-by: Boris Ranto <branto@redhat.com>
There should be no need to use sudo when writing or using these files.
It creates an issue when the user running ansible-playbook does not
have sudo privs.
Signed-off-by: Andrew Schoen <aschoen@redhat.com>
Some deployments can't copy infrastructure playbooks outside of the
infrastructure-playbooks directory. Thus they use ANSIBLE_ROLES_PATH to
overcome this. However some roles have 'playbook_dir' hardcoded, which
results in wrong path since the execution comes from
infrastructure-playbooks. Basically the role triggered by a playbook
from infrastructure-playbooks believes that the roles are in
infrastructure-playbooks/roles. This commit fixes that.
Signed-off-by: Sébastien Han <seb@redhat.com>
This will give us more flexibility and the possibility to deploy a client node
for an external ceph-cluster.
related BZ:
https://bugzilla.redhat.com/show_bug.cgi?id=1469426Fixes: #1670
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
Before this commit we were forcing ipv4 which might not be available.
Now setting ip_version to ipv4 or ipv6 will give you the right support.
Closes: https://bugzilla.redhat.com/show_bug.cgi?id=1484189
Signed-off-by: Sébastien Han <seb@redhat.com>
This commit eases the use of the
infrastructure-playbooks/switch-from-non-containerized-to-containerized-ceph-daemons.yml
playbook. We basically run it with a couple of pre-tasks and then we let
the playbook run the docker roles.
It obviously expect to have proper variables configured in order to
work.
Signed-off-by: Sébastien Han <seb@redhat.com>
The lvm_volumes variable is now a list of dictionaries that represent
each OSD you'd like to deploy using ceph-volume. Each dictionary must
have the following keys: data, journal and data_vg. Each dictionary also
can optionaly provide a journal_vg key.
The 'data' key represents the lv name used for the OSD and the 'data_vg'
key is the vg name that the given lv resides on. The 'journal' key is
either an lv, device or partition. The 'journal_vg' key is optional and
must be the vg name for the journal lv if given. This key is mainly used
for purging of the journal lv if purge-cluster.yml is run.
For example:
lvm_volumes:
- data: data_lv1
journal: journal_lv1
data_vg: vg1
journal_vg: vg2
- data: data_lv2
journal: /dev/sdc
data_vg: vg1
Signed-off-by: Andrew Schoen <aschoen@redhat.com>
Resolves issue: Multiple RGW Ceph.conf Issue #1258
In multi-RGW setup, in ceph.conf the RGW sections
contain identical bind IP in civetweb line. So this
modification fixes that issue and puts the right IP
for each RGW.
Signed-off-by: SirishaGuduru SGuduru@walmartlabs.com
Modified ceph-defaults and ran generate_group_vars_sample.sh
group_vars/osds.yml.sample and group_vars/rhcs.yml.sample are
not part of the changes. But they got modified when
generate_group_vars_sample.sh is ran to generate group_vars/
all.yml.sample.
Uncommented added variables in ceph-defaults
Updated tests by adding value for radosgw_interface
Added radosgw_interface to centos cluster tests
Modified ceph-rgw role,rebased and ran generate_group_vars_sample.sh
In ceph-rgw role removed check_mandatory_vars.yml.
Rebased on master.
Ran generate_group_vars_sample.sh and then the below files got
modified.
The rbd-mirror daemon will be HA under luminous and new daemon health
features require a way to uniquely identify rbd-mirror instances.
Signed-off-by: Jason Dillaman <dillaman@redhat.com>
To be properly evaluated the "skipped" conditions must always have the
first place on the list of condition, otherwise the other conditions are
evaluated before and make the task fail.
Closes: https://github.com/ceph/ceph-ansible/issues/1733
Signed-off-by: Sébastien Han <seb@redhat.com>
Problem: task "check for a ceph socket in containerized deployment" will
be skipped if we are not an OSD.
with_items are still evaluated before when conditions so if the task was
skipped the dict will be empty and then fail.
Adding a "not skipped" condition skips the execution of the task.
Closes: https://bugzilla.redhat.com/show_bug.cgi?id=1482061
Signed-off-by: Sébastien Han <seb@redhat.com>
This commits force ceph-common to be installed early in deployment on
nodes.
For instance, ceph-rbdmirror doesn't have the CLI installed while it is
needed for some tasks which uses it to set some facts.
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
ceph services can fail to start under certain circumstances (for
example, when running in a container) because the default systemd
service configuration causes namespace issues.
To work around this we can override the system service settings by
placing an overrides file in the ceph-<service>@.service.d directory.
This can be generic so as to allow any potential changes required to
the ceph-<service> service files.
The overrides file is only setup when the
"ceph_<service>_systemd_overrides" config_template override variable is
specified.
The available service systemd override files are as follows:
ceph_mds_systemd_overrides
ceph_mgr_systemd_overrides
ceph_mon_systemd_overrides
ceph_osd_systemd_overrides
ceph_rbd_mirror_systemd_overrides
ceph_rgw_systemd_overrides
The original fix to issue #1755 only set the permissions on
the monitors to which the key was copied, but not the original
monitor where the key was created. Thus, we use a separate task
to set the permission of the key.
The openstack_keys structure now supports a key called mode
whose value is a string that one could pass to chmod to set
the mode of the key file. The ansible file module applies the
mode to all openstack keys with this property.
Fixes: #1755
This is under the MDS role instead of the mon role because that role
does not create the filesystem under docker.
Signed-off-by: Douglas Fuller <dfuller@redhat.com>
This scenario will create OSDs using ceph-volume and is only available
in ceph releases greater than Luminous.
Signed-off-by: Andrew Schoen <aschoen@redhat.com>
There is only two main scenarios now:
* collocated: everything remains on the same device:
- data, db, wal for bluestore
- data and journal for filestore
* non-collocated: dedicated device for some of the component
Signed-off-by: Sébastien Han <seb@redhat.com>
in containerized deployment, if you try to update your `ceph.conf` file
it won't be actually updated on your nodes because it is overwritten by
the copy of the file which is present in your fetch directory.
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
Move `fsid`,`monitor_name`,`docker_exec_cmd` and `ceph_release` set_fact
to `ceph-defaults` role.
It will allow to reuse these facts without having to play `ceph-common`
or `ceph-docker-common`.
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
This will give us more flexibility and avoid a lot of useless when
skipping all tasks from a non-desired role.
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
Add a new role `ceph-defaults`.
This role aims to handle all defaults vars for `ceph-docker-common` and
`ceph-common` and set basic facts (eg. `fsid`)
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
Merge `ceph-docker-common` and `ceph-common` defaults vars in
`ceph-defaults` role.
Remove redundant variables declaration in `ceph-mon` and `ceph-osd` roles.
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
We don't want to have heterogeous ceph.conf anymore and believe that we
should have the right section for the running daemon.
If we don't do this and use profiles, e.g: rgw, we will get a new rgw
section on some of the nodes.
Signed-off-by: Sébastien Han <seb@redhat.com>
It is mandatory now to set the Ceph version you want to install, e.g:
ceph_stable_release: luminous
To find the release names, you can look at the release not doc:
http://docs.ceph.com/docs/master/release-notes/
Signed-off-by: Sébastien Han <seb@redhat.com>
ceph-disk is responsable for enabling the unit file if needed. Actually
since https://github.com/ceph/ceph/pull/12241 it seems that it's not
even needed. On an event of a restart, udev rules will be trigger and
they will ceph-disk activate the device too so the 'enabled' is not
needed.
Closes: https://github.com/ceph/ceph-ansible/issues/1142
Signed-off-by: Sébastien Han <seb@redhat.com>
OSDs get started by ceph-disk before the ceph.conf file is written
with a crush location. That results in a crush map without configured
crush location.
To prevent this, we have to restart the OSDs during the initial setup
after the crush location was added to the ceph.conf file.
We have multiple issues with ceph-disk's cli with bluestore and Ceph
releases. This is mainly due to cli changes with Luminous. Luminous
introduced a --bluestore and --filestore options which respectively does
not exist on releases older than Luminous. The default store being
bluestore on Luminous, simply checking for the store is not enough so we
have to build a specific command line for ceph-disk depending on the
Ceph version we are running and the desired osd_store.
Signed-off-by: Sébastien Han <seb@redhat.com>
The keys and openstack_keys structure now supports an optional
key called acls whose value is a list of strings one could pass
to setfacl. The ansible ACL module applies the ACLs to all
openstack keys with this property.
Fixes: #1688
Remove `rgw enable static website` and `rgw enable usage log` from
ceph.conf and make it usable with ceph_config_overrides as profiles.
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
This commit introduces a new directory called "profiles" which
contains some set of variables for a particular use case. These profiles
provide guidance for certain scenarios such as:
* configuring rgw with keystone v3
Signed-off-by: Sébastien Han <seb@redhat.com>
According to Alfredo, this was used for gitbuilders. Right now shaman/chacra
dev repos are unsigned.
Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
Move condition at task level and not at include level to make `fsid`
variable available for all roles.
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
Some tasks fetch file to `{{ fetch_directory }}/docker_mon_files` and
then try to copy from `{{ fetch_directory }}/{{ fsid }}`. That causes
the playbook to fail.
Fixes: #1683
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
To keep consistency between `{{ openstack_keys }}` and `{{ keys }}`
respectively in `ceph-mon` and `ceph-client` roles.
This commit also add the possibility to set mds caps.
Fixes: #1680
Co-Authored-by: John Fulton <johfulto@redhat.com>
Co-Authored-by: Giulio Fidente <gfidente@redhat.com>
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
The ceph.conf template needs to look for the value of monitor_interface
in hostvars[host] because there might be different values set per host.
Signed-off-by: Andrew Schoen <aschoen@redhat.com>
In Luminous, ceph-disk defaults to bluestore so all our scenarios are
using bluestore, we need to force testing both.
Signed-off-by: Sébastien Han <seb@redhat.com>
Co-Authored-by: Guillaume Abrioux <gabrioux@redhat.com>
In addition to ceph/ceph-docker@69d9aa6, this explains how to deploy a
containerized cluster with a custom admin secret.
Basically, just need to pass the `admin_secret` defined in your
`group_vars/all.yml` to the `ceph_mon_docker_extra_env` variable.
Eg:
`ceph_mon_docker_extra_env: -e CLUSTER={{ cluster }} -e FSID={{ fsid }}
-e MON_NAME={{ monitor_name }} -e ADMIN_SECRET={{ admin_secret }}`
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
Fail with a sane message if the devices or raw_journal_devices variables
are strings instead of lists during manual device assignment.
Signed-off-by: Douglas Fuller <dfuller@redhat.com>
rsync is required by the ansible synchronize package. Ensure
it is installed when local installation is selected.
Signed-off-by: Douglas Fuller <dfuller@redhat.com>
Add a new parameter `admin_secret` that allow to deploy a ceph cluster
with a custom admin secret.
Fix: #1630
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
This commits refactors how we deploy bluestore. We have existing
scenarios that we don't want to change too much. This commits eases the
user experience by now changing the way you use scenarios. Bluestore is
just a different interface to store objects but the scenarios more or
less remain the same.
If you set osd_objectstore == 'bluestore' along with
journal_collocation: true, you will get an OSD running bluestore with DB
and WAL partitions on the same device.
If you set osd_objectstore == 'bluestore' along with
raw_multi_journal: true, you will get an OSD running bluestore with a
dedicated drive for the rocksdb DB, then the remaining
drives (used with 'devices') will have WAL and DATA collocated.
If you set osd_objectstore == 'bluestore' along with
raw_multi_journal: true and declare bluestore_wal_devices you will get
an OSD running bluestore with a dedicated drive for rocksdb db, a
dedicated drive partition for rocksdb WAL and a dedicated drive for
DATA.
Signed-off-by: Sébastien Han <seb@redhat.com>
There is no need for 2 variables to enable bluestore, prior to this
patch one had to do the following to activate bluestore:
osd_objectstore: bluestore
bluestore: true
Now you just need to set `osd_objectstore: bluestore`.
Fixes: https://github.com/ceph/ceph-ansible/issues/1475
Signed-off-by: Sébastien Han <seb@redhat.com>
remove `ceph_mon_docker_interface` and use `monitor_interface` instead
for both containerized and non-containerized deployment.
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
Some variables are missing from ceph-docker-common role since the
include of check_mandatory_vars.yml has been re-added in the ceph-mon
role.
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
The check regarding the networking scenario configuration has been
moved from ceph-common to ceph-mon in 1de8176 but the include was not re-added
in 189f4fe
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
e8187f6 does not fix the ipv6 as expected since `ansible_default_*` are
filled with the IP address carried by the network interface used by the
default gateway route. By the way, it assumes that the MON_IP address will
be this IP address which is not always the case.
We need to keep using the previous fact but add some intelligence in the
template to determine how to retrieve the ipv4|ipv6 address since the path
to the fact in `hostvars` is not the same according to ipv4 vs ipv6 case.
Fix: 1569
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
Add an extra variable to the openstack pools, which creates them with
defined rules. This will allow to place different pools on e.g.
different type of disks.
This commit will also set a new default rule when defined and move
the rbd pool to the new rule.
Somehow the shell module will return an error if the command line is not
next to it.
Plus fixed the import with the right path.
Signed-off-by: Sébastien Han <seb@redhat.com>
Followup on https://github.com/ceph/ceph-ansible/pull/1469 where we
merged most of the container code from roles/ceph-*/task/docker/*.yml
into roles/ceph-docker-common/tasks/
It seems that we forgot to remove the original files.
Signed-off-by: Sébastien Han <seb@redhat.com>
The new test in the checks PGs are no longer working on distributions
where /bin/sh isn't linked to /bin/bash.
Fix: #1619
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
OpenStack's Gnocchi service expects to have a pool called "metrics".
This change addess "metrics" to the list of `openstack_pools` and
creates a corresponding key. It is only run if the user sets
`openstack_config: false`.
The current handler only restarts one OSD on each OSD server. After
the first one the handler stops, not matter what results the checks had.
Co-Authored-By: Gaudenz Steinlin (@gaudenz)
Remove "osd mkfs type" and the other pre-Bluestore parameters from the
generated ceph.conf so that disk activation on OSDs will work. The
current default xfs config results in a failed deployment and
incorrect partition metadata.
For newly created cluster the command: ceph --cluster {{ cluster }} osd
pool get rbd size does not respond properly.
We only want to check if the rbd pool exists, so we know use an ls |
grep approach.
Closes: https://github.com/ceph/ceph-ansible/issues/1547
Signed-off-by: Sébastien Han <seb@redhat.com>
Rewrite the check_pgs by using json parsing instead of complex regexp to
parse the `ceph -s` output.
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
Add a default value for `ceph_docker_on_openstack` to avoid a
conditional check error for the task `pause after docker install before starting` in
`roles/ceph-docker-common/tasks/pre_requisites/prerequisites.yml`
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
The fact ['ansible_$interface']['ipv4'] is a dictionary where
['ansible_$interface']['ipv6'] is a list. If we use
ansible_default_ipv6|ipv4 is is always a dictionary which allows us to
get the ipv6 and ipv4 address without adding more complexity to the
template.
Signed-off-by: Andrew Schoen <aschoen@redhat.com>
For some reason we changed the check of pgs but it appears it could be
dangerous because the current check might satisfied as long as 1 PG is
active+clean.
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
since `-e CEPH_DAEMON=OSD_CEPH_DISK_ACTIVATE` is already hardcoded in
`eph-osd-run.sh.j2` there is no need to add `-e
CEPH_DAEMON=OSD_CEPH_DISK_ACTIVATE` as a default value in defaults vars.
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
`ceph-docker-common`:
At the moment there is a lot of duplicated tasks in each
`./roles/ceph-<role>/tasks/docker/main.yml` that could be refactored in
`./roles/ceph-docker-common/tasks/main.yml`.
`*_containerized_deployment` variables:
All `*_containerized_deployment` have been refactored to a single
variable `containerized_deployment`
duplicate `cephx` variables in `group_vars/* have been removed.
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
We only check for everything expect 'distro' because that
is a valid way of deploying RHCS, with preprepared repos
present on the nodes.
Signed-off-by: Sébastien Han <seb@redhat.com>
Problem: we could end up in situation where we would install a package
on a machine that does not have the right repo enabled. Because the
condition was set to OR we weren't pinning a particular host but just a
condition. Let's say someone sets 'ceph_origin == "distro"', this would
try to install OSD packages on Monitors.
Solution: use a AND condition to first pin to the group_name (which
identifies a set of hosts) AND then after this one of the installation
condition.
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1453119
Co-Authored-By: https://github.com/zhsj
Signed-off-by: Sébastien Han <seb@redhat.com>
Problem: fail to deploy a containerized Ceph cluster with ipv6
Solution: do not hardcode ipv4 when bootstrapping the container.
Now use ip_version: ipv6 to get a containerized cluster deployed with
ipv6.
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1451786
Signed-off-by: Sébastien Han <seb@redhat.com>
In addition to `196fa7e` this commit check if a container has been
already launched and delete it before retrying the ceph osd prepare
process.
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
The CI on Docker is reporting the following error:
STDERR:
Error EINVAL: bad entity name
This is due to the fact that this auth entity name does not exist on
Jewel so we should not create that key when running Jewel containers.
Fixes: https://github.com/ceph/ceph-ansible/issues/1514
Signed-off-by: Sébastien Han <seb@redhat.com>
Already documented in the Red Hat Ceph Storage 2 Installation Guide
for Red Hat Enterprise Linux, but not here
Signed-off-by: Florian Klink <flokli@flokli.de>
"rgw override bucket index max shards" and
"rgw bucket default quota max objects" were in the
client section of the ceph.conf and not being
applied, this commit moves them to global
Resolves: bz#1391500
Signed-off-by: Ali Maredia <amaredia@redhat.com>
We shouldn't need this anymore as the upgrade bug that
debian_ceph_packages was used to workaround should have
been fixed as of jewel.
See https://github.com/ceph/ceph-ansible/issues/1481 for more
detailed information.
Signed-off-by: Andrew Schoen <aschoen@redhat.com>
Change civetweb_num_thread default to 100
Add capability to override number of pgs for
rgw pools.
Add ceph.conf vars to enable default bucket
object quota at users choosing into the ceph.conf.j2
template
Resolves: rhbz#1437173
Resolves: rhbz#1391500
Signed-off-by: Ali Maredia <amaredia@redhat.com>
Restore the check_socket that was removed by `5bec62b`.
This commit also improves the logging in `restart_*_daemon.sh` scripts
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
Prior to this change, ansible was only checking for the existence of the
package, now if upgrade_ceph_packages is true this means we are
performing an upgrade.
Fixes: https://bugzilla.redhat.com/show_bug.cgi?id=1442016
Signed-off-by: Sébastien Han <seb@redhat.com>
Proof-of-concept clusters or actual production clusters will never want to use this. We also do not test it anywhere for this same reason.
Signed-off-by: Gregory Meno <gmeno@redhat.com>
This is to allow ceph-mgr daemons to remote control
osd and mds daemons with MCommand messages.
Fixes: http://tracker.ceph.com/issues/19713
Signed-off-by: John Spray <john.spray@redhat.com>
Without this, we don't test the mgr role so we need to add it.
Co-Authored-by: Guillaume Abrioux <gabrioux@redhat.com>
Signed-off-by: Sébastien Han <seb@redhat.com>
This is the same fix as bc846b7da6
applied to the other part of the code-base that builds ceph.conf (I'd
missed that 349b9ab3e7 had duplicated
this code).
Signed-off-by: Matthew Vernon <mv3@sanger.ac.uk>
Ansible's assemble module by default will put all files in the src
directory together into dest. We only want to put {{ cluster }}.conf
and osd.conf together, not anything that might have found its way into
/etc/ceph/ceph.d (e.g. files left by the sysadmin taking backups
before an ansible run). So specify a regexp that matches only those
two files.
Signed-off-by: Matthew Vernon <mv3@sanger.ac.uk>
Ansible evaluates the 'with_items' before the 'when' so if the inventory
does not have the group declared it'll fail. To fix this, we set an
empty array to make the with_items happy and then evaluate with the
'when'.
Signed-off-by: Sébastien Han <seb@redhat.com>
Prior to this change we were deploying a monitor using tis fqdn name but
we were checking its state and performing actions on it using its
shortname.
Signed-off-by: Sébastien Han <seb@redhat.com>
The Ceph Manager daemon (ceph-mgr) runs alongside monitor daemons, to
provide additional monitoring and interfaces to external monitoring and
management systems.
Only works as of the Kraken release.
Co-Authored-By: Guillaume Abrioux <gabrioux@redhat.com>
Signed-off-by: Sébastien Han <seb@redhat.com>
ceph-create-keys unit file was removed here:
* 8bcb4646b6
* dc5fe8d415
As a consequence the systemctl preset command now fails to run since the
unit does not exist anymore. Due to the redirection in /dev/null we
don't know what's happening.
Ultimately the mon unit doesn't get enabled and the mon service won't
start after reboot.
Removing the old/non-existent unit makes the command succeed now.
ceph fix: https://github.com/ceph/ceph/pull/14226
Signed-off-by: WingkaiHo <sanguosfiang@163.com>
Co-Authored-By: Sébastien Han <seb@redhat.com>
Until now, only the first task were executed.
The idea here is to use `listen` statement to be able to notify multiple
handler and regroup all of them in `./handlers/main.yml` as notifying an
included handler task is not possible.
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
As reported in
https://github.com/ceph/ceph-ansible/issues/1403 when devices are held
by lvm and `osd_auto_discovery` is set to true, it's not enough to check
for a partition count = 0 since Ansible does not report.
This patch also looks for 'holders' which in a case of lvm corresponds
to the name of the pv. Now we also look for holders = 0.
Fixes: #1403
Signed-off-by: Sébastien Han <seb@redhat.com>
Problem: too many different commands to do the same thing. The 'cut'
command on infrastructure-playbooks/purge-cluster.yml was also wrong.
This sed command from osixia in ceph-docker
https://github.com/ceph/ceph-docker/pull/580/ addresses all the
scenarios.
Signed-off-by: Sébastien Han <seb@redhat.com>
ntp is still installed even if ntp_service_enabled is set to false.
That could be a problem if the time synchronization is managed by
something else than ceph-ansible or if you want to use different NTP
implementation as suggested in #1354.
Fixes: #1354
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
Signed-off-by: Guits <gabrioux@redhat.com>
If a group of hosts is empty, (for instance 'mdss', in case of a
deployment without any mds node), the playbook will fails when trying
to restart service with `"'dict object' has no attribute u'XXX'"` error.
The idea here is to force the `with_items` statements in all included handler tasks
to get at least an empty array.
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
systctl tuning should be in the sysctl.d directory. This creates
a seperation from what values were set specific to ceph, and what
values were set by the operator.
Signed-off-by: Tyler Brekke <tbrekke@redhat.com>
With ' in osd_crush_location, systemd will show this error:
ceph-osd-prestart.sh[2931]: Invalid command: invalid chars ' in 'root=
Signed-off-by: Christian Zunker <christian.zunker@codecentric.de>
This fixes issue #1299. According to @ktdreyer s comment in the ticket,
he fixed the web server config so also older (non-SNI) python clients
can use the uri module here.
After the jewel release the mon startup does not generate keys, but it's
still harmless to call ceph-create-keys with jewel because this task has
a 'creates' argument that will cause it not to run if the keys already
exist.
Removing this when condition also allows the downstream CI tests to
install kraken or luminous without resetting ceph_stable_release, which does not
pertain to rhcs.
Signed-off-by: Andrew Schoen <aschoen@redhat.com>
This is not only for monitors, but also mds, rgw and rbd mirror so
making the var name more generic:
ceph_docker_enable_centos_extra_repo
Signed-off-by: Sébastien Han <seb@redhat.com>
Sometimes the socket appears during the 5th attempt and sometimes not so
increasing the timeout a little bit.
Signed-off-by: Sébastien Han <seb@redhat.com>
Add the possibility to create openstack pools and keys even for containerized deployments
Fix: #1321
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
This removes the implicit order requirement when using OSD fragments.
When you use OSD fragments and ceph-osd role is not the last one,
the fragments get removed from ceph.conf by ceph-common.
It is not nice to have this code at two locations, but this is
necessary to prevent problems, when ceph-osd is the last role as
ceph-common gets executed before ceph-osd.
This could be prevented when ceph-common would be explicitly called
at the end of the playbook.
Signed-off-by: Christian Zunker <christian.zunker@codecentric.de>
This option was missing for rrgw, mds, rbd mirror and nfs making these
daemon impossible to run on a kv deployment with containers.
Signed-off-by: Sébastien Han <seb@redhat.com>
Prior to this change, ceph-ansible would install the main NFS Ganesha
server daemon on Ubuntu, but it would skip the Ceph FSALs.
Running "apt-get install nfs-ganesha" will only install the main NFS Ganesha
server. It does *not* pull in the RGW FSAL
(/usr/lib/x86_64-linux-gnu/ganesha/libfsalrgw.so)
Running "apt-get install nfs-ganesha-fsal" will install the RGW FSAL as
well as the main NFS Ganesha server package.
Signed-off-by: Ken Dreyer <kdreyer@redhat.com>
This patch introduces calamari_debug option which will turn on debugging
for calamari before initializing and running it.
Signed-off-by: Boris Ranto <branto@redhat.com>
From Josh Durgin, "I'd recommend not setting vfs_cache_pressure in
ceph-ansible. The syncfs issue is still there, and has caused real
problems in the past, whereas there hasn't been good data showing lower
vfs_cache_pressure is very helpful - the only cases I'm aware of have
shown it makes little difference to performance."
https://bugzilla.redhat.com/show_bug.cgi?id=1395451
Install package from official repos rather than pip when using RHEL.
This commit fix https://bugzilla.redhat.com/show_bug.cgi?id=1420855
Also this commit Refact all `roles/ceph-*/tasks/docker/pre_requisite.yml`
to avoid a lot of duplicated code.
Fix: #1303
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
This was needed for Hammer and older version, not needed anymore since
we have a 'ceph' user to run ceph processes.
Signed-off-by: Sébastien Han <seb@redhat.com>
Check if ceph filesystem already exists before creating it.
If the ceph filesystem doesn't exist, execute the task only on one node.
Fix: #1314
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
Since distro will not allow /usr/share to be writable (e.g: atomic) so
we let the operator decide where to put that script.
Signed-off-by: Sébastien Han <seb@redhat.com>
Oh yeah! This patch adds more fine grained control on how we run the
activation osd container. We now use --device to give a read, write and
mknodaccess to a specific device to be consumed by Ceph. We also use
SYS_ADMIN cap to allow mount operations, ceph-disk needs to temporary
mount the osd data directory during the activation sequence.
This patch also enables the support of dedicated journal devices when
deploying ceph-docker with ceph-ansible.
Depends on https://github.com/ceph/ceph-docker/pull/478
Signed-off-by: Sébastien Han <seb@redhat.com>
As of Infernalis, the Ceph daemons run as an unprivileged "ceph" UID,
and this is by design.
Commit f19b765 altered the default
civetweb port from 80 to 8080 with a comment in the commit log about
"until this gets solved"
Remove the comment about permissions on Infernalis, because this is
always going to be the case on the Ceph versions we support, and it
is just confusing.
If users want to expose civetweb to s3 clients using privileged TCP
ports, they can redirect traffic with iptables, or use a reverse proxy
application like HAproxy.
Signed-off-by: Andrew Schoen <aschoen@redhat.com>
This avoids a situation where during a rolling_update we try to talk to
a mon to get the fsid and if that mon is down the playbook hangs
indefinitely.
Signed-off-by: Andrew Schoen <aschoen@redhat.com>
This gives us more flexibility than installing the ceph-release package
as we can easily use different mirrors. Also, I noticed an issue when
upgrading from jewel -> kraken as the ceph-release package for those
releases both have the same version number and yum doesn't know to
update anything.
Signed-off-by: Andrew Schoen <aschoen@redhat.com>
To configure kernel the task is using "command" module which is not
respect operator ">". So this task just print to "stdout": "never >
/sys/kernel/mm/transparent_hugepage/enabled"
fix: #1319
Signed-off-by: Sébastien Han <seb@redhat.com>
Some playbooks use [0-9]*, others use \d+$
The latter is more correct since cluster name may contain numbers.
Signed-off-by: Shengjing Zhu <zsj950618@gmail.com>
So unit files were stored in /var/lib/ceph some where in
/etc/systemd/system. Now they are all under /etc/systemd/system.
closes: #1296
Signed-off-by: Sébastien Han <seb@redhat.com>
If cephx is disabled it is not necessary to include `facts_mon_fsid.yml`
in `roles/ceph-common/tasks/facts.yml`.
Fix: #1300
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
We changed the way we declare image.
Prior to this patch we must have a "user/image:tag"
format, which is incompatible with non docker-hub registry where you
usually don't have a "user". On the docker hub a "user" is also
identified as a namespace, so for Ceph the user was "ceph".
Variables have been simplified with only:
* ceph_docker_image
* ceph_docker_image_tag
1. For docker hub images: ceph_docker_name: "ceph/daemon" will give
you the 'daemon' image of the 'ceph' user.
2. For non docker hub images: ceph_docker_name: "daemon" will simply
give you the "daemon" image.
Infrastructure playbooks have been modified as well.
The file group_vars/all.docker.yml.sample has been removed as well.
It is hard to maintain since we have to generate it manually. If
you want to configure specific variables for a specific daemon simply
edit group_vars/$DAEMON.yml
Fixes: https://bugzilla.redhat.com/show_bug.cgi?id=1420207
Signed-off-by: Sébastien Han <seb@redhat.com>
We shouldn't test directly the value of
`ceph_conf_overrides.global.osd_pool_default_pg_num` because this can
cause the playbook to fail if the key `global` is not present in
`ceph_conf_overrides`. Therefore we have to use the facts that have been
defined earlier.
Fix: #1242
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
On ubntu systems mkdir is in /bin where on atomic it is /usr/bin/.
We use the shell built-in function "command" to find its right location.
Signed-off-by: Sébastien Han <seb@redhat.com>
Since we now only support systemd has an init system we can finally
treat containers as processes using systemd and this for all the
distros.
Signed-off-by: Sébastien Han <seb@redhat.com>
This commits allows us to restart Ceph daemon machine by machine instead
of restarting all the daemons in a single shot.
Rework the structure of the handler for clarity as well.
Signed-off-by: Sébastien Han <seb@redhat.com>
According to #1216, we need to simply the code by removing the
support of anything before Jewel.
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
Some users purge their environments and leave it in a non-optimal state.
e.g: packages are still installed but /etc/ceph and /var/lib/ceph don't
exist anymore. This will result in multiple failures across the play,
sometimes hard to detect. Populating these directories "just in case"
should help us solving these problems.
Closes: #1253
Signed-off-by: Sébastien Han <seb@redhat.com>
Sometimes users for testing, tend to delete the whole /var/lib/ceph and
then run ansible again, OSD will never come up if we do not create their
directory.
Signed-off-by: Sébastien Han <seb@redhat.com>
This patch makes sure we set the proper pool size on the rbd pool.
Usually during bootstrap the rbd pool size is not honoured so we need to
add this workaround.
Signed-off-by: Sébastien Han <seb@redhat.com>
This allows the user to set ip_version to either ipv4 or ipv6. This
resolves a bug where monitor_address is set to an ipv6 address, but the
template fails to render because it's hardcoded to look for an 'ipv4'
key in the ansible facts.
See: https://bugzilla.redhat.com/show_bug.cgi?id=1416010
Signed-off-by: Andrew Schoen <aschoen@redhat.com>
Resolves: bz#1416010
could have scenario where different openstack components would
use the same pool, but the logic would create the same pool
more than once
add unique filter to account for this
Allow for more operator flexibility in the `rgw frontends` setting
while maintaining backwards compatibility with the old vars. This
allows an operator to, for example, use the civetweb settings for
implementing SSL ports.
For available civetweb configuration parameters, see:
https://github.com/civetweb/civetweb/blob/master/docs/UserManual.md
It is not enough to check for the mds to exists, it actually always does
because we declare the variable. So we need to make sure that there is a
mds host.
Signed-off-by: Sébastien Han <seb@redhat.com>
Since we introduced config_overrides we removed a lot of options from
the default template. In some cases, like mds pool, openstack pools etc
we need to know the amount of PGs required. The idea here is to skip the
task if ceph_conf_overrides.global.osd_pool_default_pg_num is not define
in your `group_vars/all.yml`.
Closes: #1145
Signed-off-by: Sébastien Han <seb@redhat.com>
Co-Authored-By: Guillaume Abrioux <gabrioux@redhat.com>
This allows for the role to be used with ansible-galaxy and to fix the
include in all the meta/main.yml files in the roles.
Signed-off-by: Andrew Schoen <aschoen@redhat.com>
The libcephfs1 package was removed from ceph-common in
cb1c06901e, however it was not synced
to group_vars/all.yml.sample using the `generate_group_vars_sample.sh`
script. This fixes up the comment formatting in the ceph-common
defaults and brings the group_vars sample back into sync.
Prior to this change, a playbook run with '--tags' or '--skip-tags'
would fail, because the ceph-common role would not include the
release.yml task, and this file defines critical things like
ceph_release.
Thanks Andrew Schoen <aschoen@redhat.com> for help with the fix.
Task put initial mon keyring in mon kv store from
ceph-mon/tasks/ceph_keys.yml is failing when cephx is disabled. The root
cause is that variable monitor_keyring is not populated by any task from
deploy_monitors.yml.
Fixes: #1211
Signed-off-by: Sébastien Han <seb@redhat.com>
Prior to this patch we had several ways to runs containers, we could use
ansible's docker module on some distro and on containers distros we were
using systemd. We strongly believe threating containers as services with
systemd is the right approach so this patch generalizes to all the
distros. These days most of the distros are running systemd so it's fair
assumption.
Signed-off-by: Sébastien Han <seb@redhat.com>
Once we have our first monitor up and running we need to add it to the
monitor store as a safety measure. Just in case the local file gets
deleted and you need to add a new monitor. Now you can retrieve this key
like this:
ceph config-key get initial_mon_keyring > initial_mon_keyring.txt
Signed-off-by: Sébastien Han <seb@redhat.com>
There is no need to become root on local_action. This will event trigger
an error on some systems as it will try to run a sudo command. If the
current user does not have passwordless sudo, Ansible will fail. Anyway
using the current user is perfectly fine and no elevation privilege is
needed.
Signed-off-by: Sébastien Han <seb@redhat.com>
The Keystone v2 APIs are deprecated and scheduled to be removed in
Q release of Openstack. This adds support for configuring RGW to
use the current Keystone v3 API.
The PKI keys are used to decrypt the Keystone revocation list when
PKI tokens are used. When UUID or Fernet token providers are used in
Keystone, PKI certs may not exist, so we now accommodate this scenario
by allowing the operator to disable the PKI tasks.
Jewel added support for user/pass authentication with Keystone,
allowing deployers to disable Keystone admin token as required
for production deployments.
This implements configuration for the new RGW Keystone user/pass
authentication feature added in Jewel.
See docs here: http://docs.ceph.com/docs/master/radosgw/keystone/
Just for clarity and because we can we now show the name of the
ceph configuration file that is generated.
Signed-off-by: Sébastien Han <seb@redhat.com>
This commit solves the situation where you lost your fetch directory and
you are running ansible against an existing cluster. Since no fetch
directory is present the file containing the initial mon keyring
doesn't exist so we are generating a new one.
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
We do not need to run another condition for 'ceph_rhcs' since the
include we came from already has it, so we are already inside this
condition.
We also spell red hat entirely instead of rh and we remove capital
letters.
Signed-off-by: Sébastien Han <seb@redhat.com>
When `ceph_stable_rh_storage` is True, every cluster node should have a
`/etc/apt/preferences.d/rhcs.pref` file with the following contents:
```
Explanation: Prefer Red Hat packages
Package: *
Pin: release o=/Red Hat/
Pin-Priority: 999
```
ceph-deploy already did this when used with ice-setup, and we need to do
the same thing with the ceph-ansible stack.
Closes: #1182 and https://bugzilla.redhat.com/show_bug.cgi?id=1404515
Signed-off-by: Sébastien Han <seb@redhat.com>
Only when ceph_origin == "upstream", install_on_redhat.yml will include
redhat_ceph_repository.yml, same as debian.
In redhat_ceph_repository.yml, ceph_custom_repo will be added.
But in check_mandatory_vars.yml, ceph_origin=="upstream" can't be combined
with ceph_custom
If previous check was not run, .stdout_lines is not a valid key on the dictionary.
To get around this, use .get("stdout_lines") instead.
Also add in a default empty list
in hammer, ceph-common depended on libcephfs (indirectly, via
python-cephfs). this is no longer the case in jewel or later, so it can
be removed from debian_ceph_packages
Signed-off-by: Casey Bodley <cbodley@redhat.com>
For readibility and clarity we do not run any tasks directly in the
main.yml file. This file should only contain include, which helps us
later to apply conditionnals if we want to.
Signed-off-by: Sébastien Han <seb@redhat.com>
This commit re-uses some of the existing ceph-ansible variables for a
containirzed deployment. There is no reasons why we should add new
variables for the containerized deployment.
Signed-off-by: Sébastien Han <seb@redhat.com>
mon_group_name variable can be used to override mons group, but
this task assumes the group is always 'mons'. So we need to use
the var to find the group name instead.
Before this patch only the address for the first mon would show
in the ceph.conf even if there were multiple mons in the inventory.
Signed-off-by: Andrew Schoen <aschoen@redhat.com>
Once the monitor process starts it will also trigger `ceph-create-keys`
which will collect the admin key and bootstrap keys. We used to force
this command because we were having issues on some distros like centos
7.0 and 7.1 not triggering this. This is fixed on centos 7.2 and not an
issue on ubuntu 14.04 or 16.04 so we can remove this task. If the
monitor hangs or fails to start the playbook will fail right after at
the "wait for client.admin key exists" task after 300sec.
Closes: #1161
Signed-off-by: Sébastien Han <seb@redhat.com>
This commit solves the situation where you lost your fetch directory and
you are running ansible against an existing cluster. Since no fetch
directory is present the file containing the fsid doesn't exist so we
are creating a new one. Later the ceph.conf gets updated with a wrong
fsid which causes problems for clients and ceph processes.
Closes: #1148
Signed-off-by: Sébastien Han <seb@redhat.com>
Adding that avoids this bug:
https://github.com/ansible/ansible/issues/18206
Without that you'll get failures like:
TASK [ceph-mon : set keys permissions]
*****************************************
task path:
/home/andrewschoen/ceph-ansible/roles/ceph-mon/tasks/ceph_keys.yml:31
fatal: [mon0]: FAILED! => {"failed": true, "msg": "'dict object' has no attribute 'stdout_lines'"}
Signed-off-by: Andrew Schoen <aschoen@redhat.com>
We removed the "apache" setting for "radosgw_frontend" in
adfdf6871e.
As part of that change, we removed the final references to
ceph-extra.repo, but I failed to clean up this file itself.
Now that nothing uses this file, delete it.
This file contained the sole reference to redhat_distro_ceph_extra, so
we can drop that variable as well.
Refactor the code using 'package' module
Fix Issue #520
(However it doesn't cover all cases because some cases are not refactorable.
Ex: because of diverging packages name between distribution)
a397922 introduced a syntax error by attempting to default an unquoted
string, which causes execution failures on some ansible versions with:
Failed to template {{ ceph_rhcs_mount_path }}: Failed to template {{ ceph_stable_rh_storage_mount_path | default(/tmp/rh-storage-mount) }}: template error while templating string: unexpected '/'
libfcgi is dead upstream (http://tracker.ceph.com/issues/16784)
The RGW developers intend to remove libfcgi support entirely before the
Luminous release.
Since libfcgi gets little-to-no developer attention or testing, remove
it entirely from ceph-ansible.
Ansible task was not properly fetching OSD cluster keyring causing
the keyring to be missing when we needed to authenticate. Similarly, we
were not properly waiting on the OSD keyring to be available before
continuing.
Signed-off-by: Ivan Font <ifont@redhat.com>
- Update rolling update playbook to support containerized deployments
for mons, osds, mdss, and rgws
- Skip checking if existing cluster is running when performing a rolling
update
- Fixed bug where we were failing to start the mds container because it
was missing the admin keyring. The admin keyring was missing because
it was not being pushed from the mon host to the ansible host due to
the keyring not being available before running the copy_configs.yml
task include file. Now we forcefully wait for the admin keyring to be
generated before continuing with the copy_configs.yml task include file
- Skip pre_requisite.yml when running on atomic host. This technically
no longer requires specifying to skip tasks containing the with_pkg tag
- Add missing variables to all.docker.sample
- Misc. cleanup
Signed-off-by: Ivan Font <ifont@redhat.com>
We have a fact that detects the package manager, so we can detect if
systemd is used. Radosgw was still using some old logic from Ubuntu.
Ubuntu 16.04 now has systemd so we don't need to configure rgw as it was
running on upstart.
Signed-off-by: Sébastien Han <seb@redhat.com>
ansible 2.2 deprecates first_available_file option which is used in
the config_template module by 'generate ceph configuration file' task.
This change syncs the config_module files from their master repository
in github.com/openstack/openstack/ansible-plugins which includes the fix
2f6cac2cf6
Signed-off-by: Alberto Murillo Silva <alberto.murillo.silva@intel.com>
Even for dmcrypt we need to check the "devices" status and
"raw_journal_devices" as well so we can fix them if there is something
wrong with them.
Signed-off-by: Sébastien Han <seb@redhat.com>
Before this commit if you had set monitor_interface in your
inventory file for a specific host it would be ignored and the value
in group_vars/all would have been used.
Also, this enables support for monitor_address again as it had been
broken by previous changes to this template.
Signed-off-by: Andrew Schoen <aschoen@redhat.com>
This is done for preventing of their use-before-definition for osd scenarios checks (should be removed after a refactor has properly seperated all the checks into appropriate roles).
Signed-off-by: Eduard Egorov <eduard.egorov@icl-services.com>
Users reported that pool_default_pg_num is not honoured for the default
pool 'rbd'. So now we check the pg num value for the RBD pool and if it
does not match pool_default_pg_num then we delete and recreate it.
We also make sure the pool is empty first, just in case someone changed
the value manually and didn't reflect the change in ceph-ansible.
The only issue with this patch is that the pool ID will not be 0 anymore
but more likely 1.
Signed-off-by: Sébastien Han <seb@redhat.com>
backward compatibility for ceph-ansible version running latest code but
using variables defined before commit: 492518a2
Signed-off-by: Sébastien Han <seb@redhat.com>
ceph-fetch-keys role currently works only if cluster name is 'ceph'.
This commit allows to set custom cluster name in 'defaults' in the same
fashion as other roles do.
This RHCS version is now generally available. Default to using it.
Signed-off-by: Alfredo Deza <adeza@redhat.com>
Signed-off-by: Ken Dreyer <kdreyer@redhat.com>
Related: rhbz#1357631
By overriding the openstack_pools variable introduced by this commit, the
deployer may choose not to create some of the openstack pools, or to add
new pools which were not foreseen by ceph-ansible, e.g. for a gnocchi
storage backend.
For backwards compatibility, we keep the openstack_glance_pool,
openstack_cinder_pool, openstack_nova_pool and
openstack_cinder_backup_pool variables, although the user may now choose
to specify the pools directly as dictionary literals inside the
openstack_pools list.
This allows us to test devices set with persistent naming such as
/dev/disk/by-*
When registering devices we can use persisent (/dev/disk/by-*) or
non-persistent (/dev/sd*). Both declarations are supported by
ceph-ansible. There was just two tasks that were not compatible with
this. Since we support using partitions directly we need to test that
because the device activation will be different. To test if the device
is a partition we use a regular expression which wasn't compatible with
the persistent device naming format (/dev/disk/by-*).
This commit solves this issue by reading the path of the symlink since
devices like /dev/disk/by-* are symlinks to devices like /dev/sd*
Signed-off-by: Sébastien Han <seb@redhat.com>
For some providers (such as upcoming Linode support), some NICs may have
multiple IP addresses. (In the case of Linode, the only NIC has a public
and private IP address.) This is normally okay as we can use the
ceph.conf cluster_network and public_network variables to force the
monitor to listen on the addresses we want. However, we also need
ansible to set the correct monitor IP addresses in "mon hosts" (i.e. the
addresses the monitors will listen on!). This new monitor_address_block
setting tells ansible which IP address to use for each monitor.
Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
- Move mon_containerized_default_ceph_conf_with_kv config from ceph-mon
to ceph-common defaults as it's used in ceph-nfs
- Update conditional to generate ganesha config when not
mon_containerized_default_ceph_conf_with_kv
- Revert change to store radosgw keyring using ansible_hostname on
ansible server so that ceph-nfs can find it
- Update ceph-ceph-nfs0-rgw-user container to use ansible_hostname
variable
Signed-off-by: Ivan Font <ivan.font@redhat.com>
use the activation scenario instead of the full ceph_disk one, we
already have a task to prepare osds so we just need to activate the
device.
working for me using vagrant :)
Signed-off-by: Sébastien Han <seb@redhat.com>
There is no need to run the actions from
roles/ceph-mon/tasks/docker/create_configs.yml
on the first monitor only since the monitor deployment happens
**serially**.
Moreover with Vagrant it's useful to allow the auto creation of the
cluster fsid, so enabling the option. If this is not desired you can
still set `fsid: 9c9c0448-0551-401d-b55b-e5b3a42bae42` for example.
Signed-off-by: Sébastien Han <seb@redhat.com>
- Gather facts only for mons before processing ceph-mon role serially in
containerized playbook sample
- Updated ceph.conf in order to generate a valid ceph.conf
Signed-off-by: Ivan Font <ivan.font@redhat.com>
- Move fsal_rgw config to ceph-common, as it's shaered with ceph-rgw
- Update all.docker.sample with NFS config
- Rename fsal_rgw to nfs_obj_gw and fsal_ceph to nfs_file_gw, because
the former names mean nothing to non-Ganesha developers
Signed-off-by: Daniel Gryniewicz <dang@redhat.com>
-First install ceph into a directory with CMake
cmake -DCMAKE_INSTALL_LIBEXECDIR=/usr/lib -DWITH_SYSTEMD=ON -DCMAKE_INSTALL_PREFIX:PATH:=/usr <ceph_src_dir> && make DESTDIR=<install_dir> install/strip
-Ceph-ansible copies over the install_dir
-User can use rundep_installer.sh to install any runtime dependencies that ceph needs onto the machine from rundep
* changed s/colocation/collocation/
* declare dmcrypt variable in ceph-common so the variables check does
not fail
Signed-off-by: Sébastien Han <seb@redhat.com>
This fixes#845 for containerized deployments. We now also mount the
/etc/localtime volume in the containers in order to synchronize the host
timezone with the container timezone.
Signed-off-by: Ivan Font <ivan.font@redhat.com>
Prior to this change, each ceph cluster node would end up with several
"qemu-client-$pid.log" files owned by root. The [client] section would
capture *all* client activity (for example the "ceph health" command,
etc), not just librbd-in-qemu.
Restrict this section to libvirt clients only so that we don't generate
these spurious log files for other Ceph client traffic.
Signed-off-by: Ken Dreyer <kdreyer@redhat.com>
Deployment fails when the ``secure_cluster`` is false:
TASK [ceph-mon : secure the cluster]
*******************************************
fatal: [saceph-mon.vm.ceph.asheplyakov]: FAILED! => {"failed": true, "msg": "'dict object' has no attribute 'stdout_lines'"}
fatal: [saceph-mon2.vm.ceph.asheplyakov]: FAILED! => {"failed": true, "msg": "'dict object' has no attribute 'stdout_lines'"}
fatal: [saceph-mon3.vm.ceph.asheplyakov]: FAILED! => {"failed": true, "msg": "'dict object' has no attribute 'stdout_lines'"}
A conditional include evaluates all included tasks with the (additional)
conditional applied to every task [1]. Thus all tasks from `secure_cluster.yml'
are always evaluated (with an additional 'when: secure_cluster' condition).
The `secure the cluster' task iterates over ``ceph_pools.stdout_lines``
even if ``secure_cluster`` is false: in loops ansible applies conditional
to every item (by design) [2]. However the `collect all the pools' task
is skipped if the very same condition evaluates to false, which leaves
the ``ceph_pools`` undefined, so the `secure the cluster' task fails:
Provide the default (empty) list to avoid the problem.
[1] http://docs.ansible.com/ansible/playbooks_conditionals.html#applying-when-to-roles-and-includes
[2] http://docs.ansible.com/ansible/playbooks_conditionals.html#loops-and-conditionalsCloses: #913
Signed-off-by: Alexey Sheplyakov <asheplyakov@mirantis.com>
Update each role's task to use the respective role's username, image
name, and image tag to check if a container is already running. This was
causing false failures because we were not matching any running
containers and subsequently running checks.yml to check the status of
cluster files being left behind.
Signed-off-by: Ivan Font <ivan.font@redhat.com>
Journal size is not mandatory anymore, a default from 5GB is being
added. A simple warning message will show up if the size is set to
something below 5GB.
Signed-off-by: Sébastien Han <seb@redhat.com>
The ceph-common role fails when you run ansible with --check. Adding
always_run to a few tasks makes the check go through easier (although
it's not foolproof).
This will help if the path to the iso exists in the originating server but not
in the remote paths. This issue is not seen if using /tmp/file.iso but does
show up when using nested paths.
Signed-off-by: Alfredo Deza <adeza@redhat.com>
Resolves: rhbz#1355762
init_system was getting the value of "systemd\n"
and was later compared to be equal to "systemd"
making the wrong scripts to be executed.
Signed-off-by: Alberto Murillo <alberto.murillo.silva@intel.com>
Add the ability to use a custom repo, rather than just upstream, RHEL,
and distro. This allows ansible to be used for internal testing.
Signed-off-by: Daniel Gryniewicz <dang@redhat.com>
If the docker image cannot be retrieved we will fail this task silently
and the playbook ultimately succeeds without a successful deployment.
This change makes it so we fail the playbook immediately.
Signed-off-by: Ivan Font <ivan.font@redhat.com>
Ceph has the ability to export it's filesystem via NFS using Ganesha.
Add a ceph-nfs role that will start Ganesha and export the Ceph
filesystems.
Note that, although support is going in to export RGW via NFS, this is
not working yet.
Signed-off-by: Daniel Gryniewicz <dang@redhat.com>
The config template is in ceph-common, not in the individual roles, so
roles referencing it need to use playbook_dir, not role_path.
Signed-off-by: Daniel Gryniewicz <dang@redhat.com>
- Check for nmap being available was not running as a local_action, when the checks using nmap were
- Various fixes on Ansible 2.x now that the above is working
This causes ceph-ansible scripts to fail when targeting Centos7 machines.
Installation fails because newer ceph package dependencies provided
by ceph-release-{version}.noarch.rpm were overridden by older
package dependency versions in default distribution repositories,
due to the fact that default distribution repositories have higher
priority.
Docker makes it difficult to use images that are not on signed
registries. This is a problem for developers, who likely won't have
access to a registry with proper signed certificates.
This allows the ability to use any docker image on the machine running
vagrant/ansible. The way it works is that the image in question is
exported locally, then sent to each target box and imported there.
Signed-off-by: Daniel Gryniewicz <dang@redhat.com>
This will allow nodes to install rhcs that do
not have access to the internet.
Signed-off-by: Andrew Schoen <aschoen@redhat.com>
Resolves: rhbz#1337601
In order to align all Ansible versions, we now use the full path for the
template. We rely on `role_path` variable. Now all the tasks using
the template module have a uniform syntax.
Might fix issue raised in #483
Signed-off-by: Sébastien Han <seb@redhat.com>
Andrew Schoen
Sébastien Han
Michael Vollman
Andy McCrae
Guillaume Abrioux
Gregory Meno
Simone Caronni
Ramana Raja
Di Xu
Randy J. Martinez
Douglas Fuller
vasishta p shastry
Ali Maredia
Ken Dreyer
Zack Cerza
Attila Fazekas
Alfredo Deza
John Fulton
Ning Yao
Patrick Donnelly
jtudelag
Greg Charot
Grant Slater
Giulio Fidente
Paul Bourke
Caleb Boylan
Eduard Egorov
Christian Berendt
Markos Chandras
Major Hayden
Konstantin Shalygin
wintamute
Jan Provaznik
Yixing Yan
Arano-kai
Joe Talerico
Fabien Brachere
Michel Rode
Al Lau
Florian Haas
Logan V
Andrew Schoen
Keith Schincke
Boris Ranto
Alfredo Deza
SirishaGuduru
Jason Dillaman
Yoann Moulin
Christian Zunker
yanyx
Peter Jenkins
Austin Workman
Pascal Watteel
David Galloway
Florian Klink
Alberto Murillo
John Spray
John McEleney
Matthew Vernon
WingkaiHo
Daniel Horak
Concubidated
yuliyang
Daniel Marks
WingKai Ho
Tobias Florek
Shengjing Zhu
Chris Wells
tynorth-cisco
Musee Ullah
Davide Belloni
Casey Bodley
Jirayut Nimsaeng
Ivan Font
James Saint-Rossy
Adam Huffman
Gunwoo Gim (a.k.a. Nicho1as)
Adam Huffman
Simon Weald
9seconds
suuuper
Musee Ullah
OrFriedmann
Proskurin Kirill
Paulo Matias
Patrick Donnelly
Ivan Font
Daniel Gryniewicz
asbishop
Christoph Dwertmann
Daniel Lin
Martin Bukatovic
Alexey Sheplyakov
Chengwei Yang
Kalle Happonen
pprokop
KGoreczny
Erik Hollensbe
Austin Brown
Huamin Chen
Matthew Rees
Jim Curtis
David C Wang
Tino Schmeier
Sam Yaple