each monitor node should select another monitor which isn't itself.
Otherwise, one node in the monitor group won't set this fact and causes
failure.
Typical error:
```
TASK [create potentially missing keys (rbd and rbd-mirror) when mon is containerized] ***
task path: /home/jenkins-build/build/workspace/ceph-ansible-prs-dev-update_docker_cluster/rolling_update.yml:200
Thursday 22 November 2018 14:02:30 +0000 (0:00:07.493) 0:02:50.005 *****
fatal: [mon1]: FAILED! => {}
MSG:
The task includes an option with an undefined variable. The error was: 'dict object' has no attribute u'mon2'
```
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
(cherry picked from commit af78173584)
During an upgrade ceph won't create keys that were not existing on the
previous version. So after the upgrade of let's Jewel to Luminous, once
all the monitors have the new version they should get or create the
keys. It's ok to have the task fails, especially for the rbd-mirror
key, which only appears in Nautilus.
Fixes: https://bugzilla.redhat.com/show_bug.cgi?id=1650572
Signed-off-by: Sébastien Han <seb@redhat.com>
(cherry picked from commit 4e267bee4f)
When checking if a key exists we also have to ensure that the key exists
on the filesystem, the key can change on Ceph but still have an outdated
version on the filesystem. This solves this issue.
Signed-off-by: Sébastien Han <seb@redhat.com>
(cherry picked from commit 691f373543)
It's easier lookup a directoriy instead of the block devices,
especially because of ceph-volume and ceph-disk have a different way to
handle devices.
Signed-off-by: Sébastien Han <seb@redhat.com>
(cherry picked from commit c14f9b78ff)
Prior to this commit we were only disabling ceph-osd units, but forgot
the ceph.target which is controlling everything and will restart the
ceph-osd units at each reboot.
Now that everything gets disabled there won't be any conflicts between
the old non-container and the new container units.
Signed-off-by: Sébastien Han <seb@redhat.com>
(cherry picked from commit cd56dad9fa)
If we mask it we won't be able to start the OSD container since now the
osd container use the osd ID as a name such as: ceph-osd@0
Fixes the error: Failed to execute operation: Cannot send after transport endpoint shutdown
Signed-off-by: Sébastien Han <seb@redhat.com>
(cherry picked from commit fe1d09925a)
change default value of `radosgw_address` to keep consistency with
`monitor_address`.
Moreover, `ceph-validate` checks if the value is '0.0.0.0' to determine
if it has to run `check_eth_rgw.yml`.
Closes: https://bugzilla.redhat.com/show_bug.cgi?id=1600227
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
(cherry picked from commit e4869ac8bd)
Adding this rule on the hypervisor will allow cluster to talk to each
other.
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
(cherry picked from commit 96ce8761ba)
setting this setting to 1 makes the CI covering the related code in the
playbook without breaking the upgrade scenarios.
Those scenarios were broken because there is a check `TASK [waiting for
clean pgs...]` in rolling_update.yml, since the pool size for
`cephfs_metadata` and `cephfs_data` are updated to `2` in
`ceph-override.json` and there is not enough osd to honor this size,
some PGs are degraded and make the mentioned check failing.
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
(cherry picked from commit 3ac6619fb9)
since `ceph-volume` introduction, there is no need to split those tasks.
Let's refact this part of the code so it's clearer.
By the way, this was breaking rolling_update.yml when `openstack_config:
true` playbook because nothing ensured OSDs were started in ceph-osd role (In
`openstack_config.yml` there is a check ensuring all OSD are UP which was
obviously failing) and resulted with OSDs on the last OSD node not started
anyway.
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
(cherry picked from commit f7fcc012e9)
when upgrading from RHCS 2.5 to 3.2, it fails because the task `create
ceph mgr keyring(s) when mon is containerized` has a when condition
`inventory_hostname == groups[mon_group_name]|last`.
First, this is incorrect because `inventory_hostname` is referring to a
mgr node, it means this condition would have never been satisfied.
Then, this condition + `serial: 1` makes the mgr keyring creating skipped on
the first node. Further, the `ceph-mgr` role tries to copy the mgr
keyring (it's not aware we are running `serial: 1`) this leads to a
failure like the following:
```
TASK [ceph-mgr : copy ceph keyring(s) if needed] ***************************************************************************************************************************************************************************************************************************************************************************
task path: /usr/share/ceph-ansible/roles/ceph-mgr/tasks/common.yml:10
Tuesday 27 November 2018 12:03:34 +0000 (0:00:00.296) 0:11:01.290 ******
An exception occurred during task execution. To see the full traceback, use -vvv. The error was: AnsibleFileNotFound: Could not find or access '~/ceph-ansible-keys/48d78ac1-e0d6-4e35-ab3e-772aea7828fc//etc/ceph/local.mgr.magna021.keyring'
failed: [magna021] (item={u'dest': u'/var/lib/ceph/mgr/local-magna021/keyring', u'name': u'/etc/ceph/local.mgr.magna021.keyring', u'copy_key': True}) => {"changed": false, "item": {"copy_key": true, "dest": "/var/lib/ceph/mgr/local-magna021/keyring", "name": "/etc/ceph/local.mgr.magna021.keyring"}, "msg": "Could not find or access '~/ceph-ansible-keys/48d78ac1-e0d6-4e35-ab3e-772aea7828fc//etc/ceph/local.mgr.magna021.keyring'"}
```
The ceph_key module is idempotent, so there is no need to have such a
condition.
Closes: https://bugzilla.redhat.com/show_bug.cgi?id=1649957
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
(cherry picked from commit 73287f91bc)
we must apply this playbook before deploying the secondary cluster.
Otherwise, there will be a mismatch between the two deployed cluster.
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
(cherry picked from commit 3d8f4e6304)
This will tremendously help debugging daemons that fail on restart by
showing the systemd unit logs.
Signed-off-by: Sébastien Han <seb@redhat.com>
(cherry picked from commit a9b337ba66)
If you deploy with 2 HDDs and 1 SDD then each subsequent deploy both
HDD drives will be filtered out, because they're already used by ceph.
ceph-volume will report this as a 'strategy change' because the device
list went from a mixed type of HDD and SDD to a single type of only SDD.
This situation results in a non-zero exit code from ceph-volume. We want
to handle this situation gracefully and report that nothing will be changed.
A similar json structure to what would have been given by ceph-volume is
returned in the 'stdout' key.
Fixes: https://bugzilla.redhat.com/show_bug.cgi?id=1650306
Signed-off-by: Andrew Schoen <aschoen@redhat.com>
(cherry picked from commit e13f32c1c5)
ceph.conf doesn't accept float value.
Typical error seen:
```
$ sudo ceph daemon osd.2 config get osd_memory_target
Can't get admin socket path: unable to get conf option admin_socket for osd.2:
parse error setting 'osd_memory_target' to '7823740108,8' (strict_si_cast:
unit prefix not recognized)
```
This commit ensures the value inserted in ceph.conf will be an integer.
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
(cherry picked from commit 68dde424f6)
if firewalld.service systemd unit is masked, the handler will fail when
trying to restart it.
Closes: https://bugzilla.redhat.com/show_bug.cgi?id=1650281
(cherry picked from commit 63b9835cbb)
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
* The default value of osd_memory_target used by ceph is 4294967296 bytes,
so use the same as ceph-ansible default.
* Convert ansible_memtotal_mb to bytes to calculate osd_memory_target
Signed-off-by: Neha Ojha <nojha@redhat.com>
(cherry picked from commit 10538e9a23)
Although this is not officially supported, this commit allows
`stable-3.2` to run against ansible 2.4.
This should ease the transition in RHOSP.
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
Don't hard code api port because it might be overridden by the user.
Signed-off-by: Mike Christie <mchristi@redhat.com>
(cherry picked from commit e2f1f81de4)
The firewall setup for igw is not getting setup because iscsi_group_name
does not it exist. It should be iscsi_gw_group_name.
Signed-off-by: Mike Christie <mchristi@redhat.com>
(cherry picked from commit a4ff52842c)
The default igw api port is 5000 in the manual setup docs and
ceph-iscsi-config package so this syncs up ansible.
Signed-off-by: Mike Christie <mchristi@redhat.com>
(cherry picked from commit a10853c5f8)
ceph-validate used to throw error for setting flags as 'true' or 'false' for True and False
Now user can set the flags 'dmcrypt' and 'osd_auto_discovery' as 'true' or 'false'
Will fix - Bug 1638325
Signed-off-by: VasishtaShastry <vipin.indiasmg@gmail.com>
(cherry picked from commit 098f42f233)
For apt-get, purge command needs to be used, instead of remove command,
to remove related configuration files. Otherwise, packages might be
shown as installed while running dpkg command even after removing them.
Fixes: https://bugzilla.redhat.com/show_bug.cgi?id=1640061
Signed-off-by: Rishabh Dave <ridave@redhat.com>
(cherry picked from commit 640cad3fd8)
When the iscsi purge playbook is run we stop the gw and api daemons but
not tcmu-runner which I forgot on the previous PR.
Fixes Red Hat BZ:
https://bugzilla.redhat.com/show_bug.cgi?id=1621255
Signed-off-by: Mike Christie <mchristi@redhat.com>
(cherry picked from commit b523a44a1a)
Adding more memory to VMs for rgw_multisite scenarios could avoid this error
I have recently hit in the CI:
(It is worth it to set 1024Mb since there is only 2 nodes in those
scenarios.)
```
fatal: [osd0]: FAILED! => {
"changed": false,
"cmd": [
"docker",
"run",
"--rm",
"--entrypoint",
"/usr/bin/ceph",
"docker.io/ceph/daemon:latest-luminous",
"--version"
],
"delta": "0:00:04.799084",
"end": "2018-10-29 17:10:39.136602",
"rc": 1,
"start": "2018-10-29 17:10:34.337518"
}
STDERR:
Traceback (most recent call last):
File "/usr/bin/ceph", line 125, in <module>
import rados
ImportError: libceph-common.so.0: cannot map zero-fill pages: Cannot allocate memory
```
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
run commands on containers when containerized deployments.
(At the moment, all commands are run on the host only)
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
Add a playbook that will upload a file on the master then try to get
info from the secondary node, this way we can check if the replication
is ok.
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
This will setup 2 cluster with rgw multisite enabled.
First cluster will act as the 'master', the 2nd will be the secondary
one.
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
since `rgw_multisite_endpoint_addr` has a default value to
`{{ ansible_fqdn }}`, it shouldn't be mandatory to set this variable.
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
- updated README-MULTISITE
- re-added destroy.yml
- added tasks in ceph-validate to make sure the
rgw multisite vars are set
Signed-off-by: Ali Maredia <amaredia@redhat.com>
We should give users the possibility to set the IP they want as
multisite endpoint, setting the default value to `{{ ansible_fqdn }}` to
not force them to set this variable.
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
- remove destroy tasks
- cleanup conditionals and syntax
- remove unnecessary realm pulls
- enable multisite to be tested in automated
testing infra
- add multisite related vars to main.yml and
group_vars
- update README-MULTISITE
- ensure all `radosgw-admin` commands are being run
on a mon
Signed-off-by: Ali Maredia <amaredia@redhat.com>
This instructs Travis to notify Galaxy when a build completes. Since 3.0
the ansible-galaxy has the ability to build and push roles from repos
with multiple roles.
Closes: https://github.com/ceph/ceph-ansible/issues/3165
Signed-off-by: Sébastien Han <seb@redhat.com>
osds-per-device needs to be passed to run_command as a string.
Otherwise, expandvars method will try to iterate over an integer.
Signed-off-by: Maciej Naruszewicz <maciej.naruszewicz@intel.com>
We do not use @<device> anymore so we don't need to perform the
readlink check anymore.
Also we are making an exception for ooo which is still using ceph-disk.
Signed-off-by: Sébastien Han <seb@redhat.com>
ceph-disk is now deprecated in ceph-ansible so let's convert all the ci
tests to use lvm instead of ceph-disk.
Signed-off-by: Sébastien Han <seb@redhat.com>
Guillaume Abrioux
Sébastien Han
Andrew Schoen
Neha Ojha
Jason Dillaman
Mike Christie
VasishtaShastry
Rishabh Dave
Ali Maredia
Maciej Naruszewicz