ceph-volume can generate large logs at some point.
debug logs by definition should be enabled only when debugging.
Let's make it customizable with a variable which is set to `False` by
default.
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
In case of failure, the systemd ExecStop isn't executed so the container
isn't removed. After a reboot of a failed node, the container doesn't
start because the old container is still present in created state.
We should always try to remove the container in ExecStartPre for this
situation.
A normal reboot doesn't trigger this issue and this also doesn't affect
nodes running containers via docker.
This behaviour was introduced by d43769d.
Closes: https://bugzilla.redhat.com/show_bug.cgi?id=1858865
Signed-off-by: Dimitri Savineau <dsavinea@redhat.com>
This changes the way we are running the podman containers via systemd.
They are now in dettached mode and Type/PIDFile set.
Closes: https://bugzilla.redhat.com/show_bug.cgi?id=1834974
Signed-off-by: Dimitri Savineau <dsavinea@redhat.com>
When using docker container engine then the systemd unit scripts only
use a dependency on the docker daemon via the After parameter.
But if docker is restarted on a live system then the ceph systemd units
should wait for the docker daemon to be fully restarted.
Closes: https://bugzilla.redhat.com/show_bug.cgi?id=1846830
Signed-off-by: Dimitri Savineau <dsavinea@redhat.com>
Since we only have one scenario since nautilus then we can just move
the container start command from ceph-osd-run.sh to the systemd unit
service.
As a result, the ceph-osd-run.sh.j2 template and the
ceph_osd_docker_run_script_path variable are removed.
Signed-off-by: Dimitri Savineau <dsavinea@redhat.com>
We shouldn't set this flag when running switch_to_containers playbook.
Otherwise the playbook fails waiting for pgs to be clean.
Closes: https://bugzilla.redhat.com/show_bug.cgi?id=1843569
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
This commits calls the `ceph_pool` module for creating ceph pools
everywhere it's needed in the playbook.
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
The condition on this task is wrong, we have to check whether
`target_size_ratio` is set in the pool definition instead.
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
When `rule_name` isn't set in `crush_rules` the osd pool creation will
fail.
This commit adds a new fact `ceph_osd_pool_default_crush_rule_name` with
the default crush rule name.
Closes: https://bugzilla.redhat.com/show_bug.cgi?id=1817586
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
Creating crush rules even with no crush hierarchy configuration is a
valid scenario so we shouldn't be bound to the first task result (which
configure crush hierarchy) to be able to add new crush rules.
Closes: https://bugzilla.redhat.com/show_bug.cgi?id=1816989
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
Since Ceph Octopus is python3 only we don't need to specify the max open
files anymore with the container engine.
Signed-off-by: Dimitri Savineau <dsavinea@redhat.com>
Since [1] we can't use osd pool without replicas (size: 1) by default.
We now need to set the mon_allow_pool_size_one flag to true in the ceph
configuration and add the --yes-i-really-mean-it flag to the osd pool
set size cli.
[1] https://github.com/ceph/ceph/commit/21508bd
Signed-off-by: Dimitri Savineau <dsavinea@redhat.com>
This commit adds condition in order to not try to customize pools size
when its type is erasure.
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
This commit adds the pg autoscaler support.
The structure for pool definition has now two additional attributes
`pg_autoscale_mode` and `target_size_ratio`, eg:
```
test:
name: "test"
pg_num: "{{ osd_pool_default_pg_num }}"
pgp_num: "{{ osd_pool_default_pg_num }}"
rule_name: "replicated_rule"
application: "rbd"
type: 1
erasure_profile: ""
expected_num_objects: ""
size: "{{ osd_pool_default_size }}"
min_size: "{{ osd_pool_default_min_size }}"
pg_autoscale_mode: False
target_size_ratio": 0.1
```
when `pg_autoscale_mode` is `True` user has to set a decent value in
`target_size_ratio`.
Given that it's a new feature, it's still disabled by default.
Closes: https://bugzilla.redhat.com/show_bug.cgi?id=1782253
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
Currently, the command executed is wrong, eg:
```
cmd:
- podman
- exec
- ceph-mon-controller-0
- ceph
- --cluster
- ceph
- osd
- pool
- create
- volumes
- '32'
- '32'
- replicated_rule
- '1'
delta: '0:00:01.625525'
end: '2020-02-27 16:41:05.232705'
item:
```
From documentation, the osd pool creation command is :
```
ceph osd pool create {pool-name} {pg-num} [{pgp-num}] [replicated] \
[crush-rule-name] [expected-num-objects]
ceph osd pool create {pool-name} {pg-num} {pgp-num} erasure \
[erasure-code-profile] [crush-rule-name] [expected_num_objects]
```
it means we pass '1' (from item.type) as value for
`expected_num_objects` by default which is very likely not what we want.
Also, this commit modifies the default value when no `rule_name` is set
to use the existing variable `osd_pool_default_crush_rule`
Closes: https://bugzilla.redhat.com/show_bug.cgi?id=1808495
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
Since ed36a11 we move the crush rules creation code from the ceph-mon to
the ceph-osd role.
To keep the backward compatibility we kept the possibility to set the
crush variables on the mons side but we didn't move the default values.
As a result, when using crush_rule_config set to true and wanted to use
the default values for crush_rules then the crush rule ansible task
creation will fail.
"msg": "'ansible.vars.hostvars.HostVarsVars object' has no attribute
'crush_rules'"
This patch move the default crush variables from ceph-mon to ceph-osd
role but also use those default values when nothing is defined on the
mons side.
Closes: https://bugzilla.redhat.com/show_bug.cgi?id=1798864
Signed-off-by: Dimitri Savineau <dsavinea@redhat.com>
Because we are relying on docker|podman for managing containers then we
don't need systemd to manage the process (like kill).
Signed-off-by: Dimitri Savineau <dsavinea@redhat.com>
We must call `container_exec_cmd` from the right monitor node otherwise
the value of the fact might mistmatch between the delegated node and the
node being played.
Closes: https://bugzilla.redhat.com/show_bug.cgi?id=1794900
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
Since this variable makes it possible to customize the mode for ceph
directories, let's make it a bit more explicit by adding a default value
in ceph-defaults.
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
Because we need to manage legacy ceph-disk based OSD with ceph-volume
then we need a way to know the osd_objectstore in the container.
This was done like this previously with ceph-disk so we should also
do it with ceph-volume.
Note that this won't have any impact for ceph-volume lvm based OSD.
Rename docker_env_args fact to container_env_args and move the container
condition on the include_tasks call.
Remove OSD_DMCRYPT env variable from the ceph-osd template because it's
now included in the container_env_args variable.
Closes: https://bugzilla.redhat.com/show_bug.cgi?id=1792122
Signed-off-by: Dimitri Savineau <dsavinea@redhat.com>
When unsetting the noup flag, we must call container_exec_cmd from the
delegated node (first mon member)
Also, adding a `run_once: true` because this task needs to be run only 1
time.
Closes: https://bugzilla.redhat.com/show_bug.cgi?id=1792320
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
When using docker 1.13.1, the current condition:
```
{% if (container_binary == 'docker' and ceph_docker_version.split('.')[0] is version_compare('13', '>=')) or container_binary == 'podman' -%}
```
is wrong because it compares the first digit (1) whereas it should
compare the second one.
It means we always use `--cpu-quota` although documentation recommend
using `--cpus` when docker version is 1.13.1 or higher.
From the doc:
> --cpu-quota=<value> Impose a CPU CFS quota on the container. The number of
> microseconds per --cpu-period that the container is limited to before
> throttled. As such acting as the effective ceiling.
> If you use Docker 1.13 or higher, use --cpus instead.
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
This commit refact the condition in the loop of that task so all
potential osd ids found are well started.
Closes: https://bugzilla.redhat.com/show_bug.cgi?id=1790212
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
This commit lets add-osd.yml in place but mark the deprecation of the
playbook.
Scaling up OSDs is now possible using --limit
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
Before this patch, the lvm2 package installation was done during the
ceph-osd role.
However we were running ceph-volume command in the ceph-config role
before ceph-osd. If lvm2 wasn't installed then the ceph-volume command
fails:
error checking path "/run/lock/lvm": stat /run/lock/lvm: no such file or
directory
This wasn't visible before because lvm2 was automatically installed as
docker dependency but it's not the same for podman on CentOS 8.
Signed-off-by: Dimitri Savineau <dsavinea@redhat.com>
cf8c6a3 moves the 'wait for all osds' task from openstack_config to the
main tasks list.
But the openstack_config code was executed only on the last OSD node.
We don't need to do this check on all OSD node so we need to add set
run_once to true on that task.
Signed-off-by: Dimitri Savineau <dsavinea@redhat.com>
When creating crush rules with device class parameter we need to be sure
that all OSDs are up and running because the device class list is
is populated with this information.
This is now enable for all scenario not openstack_config only.
Signed-off-by: Dimitri Savineau <dsavinea@redhat.com>
This adds device class support to crush rules when using the class key
in the rule dict via the create-replicated sub command.
If the class key isn't specified then we use the create-simple sub
command for backward compatibility.
Closes: https://bugzilla.redhat.com/show_bug.cgi?id=1636508
Signed-off-by: Dimitri Savineau <dsavinea@redhat.com>
If we want to create crush rules with the create-replicated sub command
and device class then we need to have the OSD created before the crush
rules otherwise the device classes won't exist.
Signed-off-by: Dimitri Savineau <dsavinea@redhat.com>
[1] introduced a regression on the fs.aio-max-nr sysctl value condition.
The enable key isn't a boolean but a string because the expression isn't
evaluated.
This string output "(osd_objectstore == 'bluestore')" is always true
because item.enable condition only matches non empty string. So the
sysctl value was applyied for both filestore and bluestore backend.
[2] added the bool filter to the condition but the filter always returns
false on string and the sysctl wasn't applyed at all.
This commit fixes the enable key value by evaluating the value instead
of using the string.
[1] https://github.com/ceph/ceph-ansible/commit/08a2b58
[2] https://github.com/ceph/ceph-ansible/commit/ab54fe2
Signed-off-by: Dimitri Savineau <dsavinea@redhat.com>
Even if this improves ceph-disk/ceph-volume performances then it also
impact the ceph-osd process.
The ceph-osd process shouldn't use 1024:4096 value for the max open
files.
Removing the ulimit option from the container engine and doing this kind
of change on the container side [1].
[1] https://github.com/ceph/ceph-container/pull/1497
Closes: https://bugzilla.redhat.com/show_bug.cgi?id=1702285
Signed-off-by: Dimitri Savineau <dsavinea@redhat.com>
[303] mktemp used in place of tempfile module
[602] Don't compare to empty string
[701] No 'galaxy_info' found
[702] Use 'galaxy_tags' rather than 'categories'
This patch also changes the ansible log_path value via the
ANSIBLE_LOG_PATH environment variable in the travis configuration to
avoid warnings.
[WARNING]: log file at /home/travis/ansible/ansible.log is not writeable
and we cannot create it, aborting
Signed-off-by: Dimitri Savineau <dsavinea@redhat.com>
There is no need to get n * number of nodes the different keyrings.
Adding a `run_once: true` here avoid running a ceph command too many
times which could be impacting large cluster deployment.
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
This commit isolates the systemd unit files generation for containers into
separate yml files in order to be able importing each corresponding roles
without playing all tasks.
This is needed so we can run ceph-ansible to render systemd unit files
so they call podman instead of docker.
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
This commit adds the `wal_devices` option support to the
ceph_volume module.
passing a devices list in `bluestore_wal_devices` will make ceph-volume
creating 1 vg using these devices to create block.wal partitions.
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
This commit adds the `block_db_devices` option support to the
ceph_volume module.
passing a devices list in `dedicated_devices` will make ceph-volume
creating 1 vg using these devices to create block.db partitions for data
devices.
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
This commit makes it possible to parametrize the ceph directories modes.
So it changes hardocded mode for ceph related directories from 0755 to
customizable with `ceph_directories_mode` variable.
Closes: #2920
Signed-off-by: Artur Fijalkowski <artur.fijalkowski@ing.com>
Co-authored-by: Guillaume Abrioux <gabrioux@redhat.com>
This commit fixes the error [306]:
`[306] Shells that use pipes should set the pipefail option`
using `/bin/bash` as executable because Debian/Ubuntu systems use `dash`
by default which doesn't have the `-o pipefail`. (See:
https://github.com/ansible/ansible-lint/issues/497#issue-424623501)
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
On containerized deployment, the OSD entrypoint runs some ceph-volume
commands (lvm/simple scan and/or activate) which perform badly without
the ulimit option.
This option was added for all previous ceph-volume commands but not on
the ceph-osd container startup.
Also updating hard limit value to 4096 to reflect default baremetal
value.
Signed-off-by: Dimitri Savineau <dsavinea@redhat.com>
just like `ceph_osd_pool_default_size`, a pool size might change after an
initial deployment. Having this condition prevents from customizing the
pool in that case.
This is not needed so let's remove it.
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
let's use `until` instead of doing test in bash using python oneliner
also, use `command` instead of `shell`.
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
the data structure has changed in octopus.
eg: the path to `num_osds` is now `["osdmap"]["num_osds"]`.
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
Guillaume Abrioux
Dimitri Savineau
Artur Fijalkowski