When configuring grafana/prometheus embed in the mgr/dashboard, we need
to use the address of the grafana-server node and not the current
hostname because mgr/dashboard and grafana/prometheus could be present
on different hosts.
We should instead rely on the grafana_server_addr variable and remove
the dashboard_url.
We don't need to execute the ceph-dashboard role on the nodes present
in the grafana-server group. This one is dedicated to the grafana and
prometheus stack.
The ceph-dashboard needs to executed where the ceph-mgr is running. It
is either on the dedicated mgr nodes or if mgr and mon are collocated
implicitly on the mon nodes.
Because we need to execute commands from a monitor node (the first one
in the mons list) we are using delegate_to option.
If there's multiple nodes running the ceph-dashboard role then the
delegated task will be executed multiple times.
Also remove a mgr config-key option not present for nautilus+ releases.
Johannes Kastl [Sat, 27 Jul 2019 14:09:26 +0000 (16:09 +0200)]
only support openSUSE Leap 15.x, fail on 42.x
openSUSE switched from 'openSUSE 13.x' to 'openSUSE Leap 42.x' and then to
'openSUSE Leap 15.x' to align with SLES15 development.
The previous logic did not correctly allow the current release, as 15.x matched
the 'less than 42.3' condition.
For now only support openSUSE Leap 15.x, and extend support once 16.x is
released (or whatever the exact version will be)
Signed-off-by: Johannes Kastl <kastl@b1-systems.de>
We don't have a reason to not apply firewall rules on the host when
using a containerized deployment.
The TripleO environments already manage the ceph firewall rules outside
ceph-ansible and set the configure_firewall variable to false.
We don't need to create a grafana system user (in fact we even don't
set the righ uid to this user) because we're using a container setup.
Instead we just need to be sure to set the owner/group to 472 (grafana
user/group from the container) like we do for ceph/167.
We don't need to set the user/group recursively on /etc/grafana
directory in a dedicated task.
Also on Ubuntu system, the ceph-grafana-dashboards isn't present so on
non containerized deployment we won't have the
/etc/grafana/dashboards/ceph-dashboard directory present (coming with
the package) so we need to be sure it exists.
The shrink_rgw scenario has been merge just after the PR about enable
ceph dashboard by default.
So right now the shrink_rgw scenrio doesn't have nodes in the grafana
group and fails.
We just need to set dashboard_enabled to false.
there's no dedicated nodes for mgr, let's use monitor nodes.
The mgr0 instance spawned isn't used, so if this node is part of the
inventory for this scenario, testinfra will complain because there's no
ceph.conf on this node.
When creating OpenStack pools, we only check if the return code from
the pool list command isn't 0 (ie: if it doesn't exist). In that case,
the return code will be 2. That's why the next condition is rc != 0 for
the pool creation.
But in containerized deployment, the return code could be different if
there's a failure on the container engine command (like container not
running). In that case, the return code could but either 1 (docker) or
125 (podman) so we should fail at this point and not in the next tasks.
According to this comment [1], this seems to be needed to detect wifi
devices.
In node exporter we can see this:
```
--collector.wifi Enable the wifi collector (default: disabled).
```
since it's enabled by default and we don't even change this in our
systemd templates for node-exporter, we can easily assume in the end
it's not needed. Therefore, let's remove this.
Dimitri Savineau [Fri, 28 Jun 2019 14:39:38 +0000 (10:39 -0400)]
dashboard: move code into a dedicated playbook
Move dashboard, grafana/prometheus and node-exporter plays into a
dedicated playbook in infrastructure-playbook directory.
To avoid using 'dashboard_enabled | bool' condition multiple time
in the main playbook we can just import the dashboard playbook or
not.
This patch also allows to use an unique dashboard playbook for
both baremetal and container playbooks.
Avoid to setup provisioners in a fully containerized environment
This commit adds a when clause to avoid the setup of grafana
provisioners in a fully containerized scenario.
This is needed when the ceph-grafana-dashboards package is not
installed and this task could result in a wrong grafana
configuration that let the container crash.
fbf4ed42aee8fa5fd18c4c289cbb80ffeda8f72e introduced a bug when
container binary is podman.
podman doesn't support ps -f using regular expression, the container id
is never set in the restart script causing the handler to fail.
The ceph nodes couldn't have the python six library installed which
could lead to error during the ceph_volume custom module execution.
ImportError: No module named six
The six library isn't useful in this module if we're sure that all
action variables passed to the build_ceph_volume_cmd function are a
list and not a string.
validate: fail if gpt header found on unprepared devices
ceph-volume will complain if gpt headers are found on devices.
This commit checks whether a gpt header is present on devices passed in
`devices` variable and fail early.
The current port value for alertmanager, grafana, node-exporter and
prometheus is hardcoded in the roles so it's not possible to change the
port binding of those services.
Both ntp and chrony daemon use variable for the service name because it
could be different depending on the GNU/Linux distribution.
This has been update in 9d88d3199 for chrony but only for the start part
not for the handler.
The commit fixes this for both ntp and chrony.
The Prometheus porrt 9090 isn't open in the firewall configuration.
Also the dashboard task on the grafana node was not required because
it's already present on the mgr node.
validate: improve message printed in check_devices.yml
The message prints the whole content of the registered variable in the
playbook, this is not needed and makes the message pretty unclear and
unreadable.
```
"msg": "{'_ansible_parsed': True, 'changed': False, '_ansible_no_log': False, u'err': u'Error: Could not stat device /dev/sdf - No such file or directory.\\n', 'item': u'/dev/sdf', '_ansible_item_result': True, u'failed': False, '_ansible_item_label': u'/dev/sdf', u'msg': u\"Error while getting device information with parted script: '/sbin/parted -s -m /dev/sdf -- unit 'MiB' print'\", u'rc': 1, u'invocation': {u'module_args': {u'part_start': u'0%', u'part_end': u'100%', u'name': None, u'align': u'optimal', u'number': None, u'label': u'msdos', u'state': u'info', u'part_type': u'primary', u'flags': None, u'device': u'/dev/sdf', u'unit': u'MiB'}}, 'failed_when_result': False, '_ansible_ignore_errors': None, u'out': u''} is not a block special file!"
```
- Remove gateway_keyring from the configuration file because it's
not used in ceph-iscsi 3.x release.
- Use config_template instead of template module for iscsi-gateway
configuration file. Because the file is an ini file and we might want
to override more parameters than those present in ceph-ansible.
- Because we can now set the pool name in the configuration, we should
use a variable for that. This is refact with the iscsi_pool_* variables
also used to configure the pool size.
Some dashboard_rgw_api_* variables are using the bool filter but those
variables are strings with an empty string as default value.
So we should test the variable against an empty string instead of a
bool.
Boris Ranto [Tue, 9 Jul 2019 20:32:38 +0000 (22:32 +0200)]
dashboard: Use upstream default port
We are currently using incorrect dashboard default port. The upstream
uses 8443 instead of 8234 by default. This should get us closer to the
upstream project.
c90f605b5 introduces the default ceph cluster name value in the rgw
socket path for the rgw restart script. But this should use the
`cluster` variable instead.
This commit also fixes this in the osd restart script.
Add package-install tag on ceph-grafana-dashboard pkg install.
According to the OSP pattern, we need the package-install tag
to control what is installed on the host. This commit just add
the missing tag to meet the TripleO requirements.
On containerized deployment we need to bind mount the ceph-iscsi
directory to avoid writing the logs in the container.
The /var/log/ceph directory isn't use by rbd-targe-api/gw services
because they have their own log directories.
Mike Christie [Thu, 30 May 2019 16:17:09 +0000 (11:17 -0500)]
igw: Add check for missing iqn
If the user is still using the older packages and does not setup
the target iqn you will just get a vague error message later on.
This adds a check during the validate task, so it is clear to the
user.
Signed-off-by: Mike Christie <mchristi@redhat.com>
Mike Christie [Mon, 13 May 2019 17:59:27 +0000 (12:59 -0500)]
igw: Add check for mismatch ceph-iscsi and iscsigws.yml settings
If the user has manually installed ceph-iscsi but is trying to setup a
iscsi object in iscsigws.yml you will just a python crash. This patch
adds a check and more user friendly error message for the case.
Signed-off-by: Mike Christie <mchristi@redhat.com>
Mike Christie [Thu, 6 Jun 2019 14:47:56 +0000 (09:47 -0500)]
igw: Update tests to use ceph-iscsi package
gateway_ip_list is depreciated and is only used when using the old
ceph-iscsi-config/cli packages that are no longer being developed
(GH repos are archived). Because ceph-iscsi-config/cli is no longer
being worked on, this modifies the tests to stress the ceph-iscsi
based installs.
Signed-off-by: Mike Christie <mchristi@redhat.com>
Mike Christie [Thu, 30 May 2019 16:28:34 +0000 (11:28 -0500)]
igw: Support ceph-iscsi package for install
This adds support for the ceph-iscsi package during install. ceph-iscsi
does not support setting up targets/gws, luns and clients with the
current library/igw_* code. Going forward those tasks should be done with
gwcli or dashboard. ceph-iscsi will only be used if the user has no iscsi
objects setup so we do not break existing setups.
The next patch will update the iscsigws.yml.sample to document that
users must not setup any iscsi object if they want to use the new
package and tools.
Signed-off-by: Mike Christie <mchristi@redhat.com>
Mike Christie [Sat, 11 May 2019 22:23:37 +0000 (17:23 -0500)]
igw: Support new ceph-iscsi package during purge
The ceph-iscsi-config and ceph-iscsi-cli packages were combined into
ceph-iscsi and its APIs changed. This fixes up the iscsi purge task to
support the new API and old one.
Signed-off-by: Mike Christie <mchristi@redhat.com>
The radosgw restart script doesn't handle this and could fail during
an upgrade.
If the SOCKETS variable isn't defined in the script then the test
command won't fail because the return code is 0
$ test -S
$ echo $?
0
There multiple issues in that script:
- The default SOCKETS value isn't defined due to a typo
SOCKET vs SOCKETS.
- Because the socket name uses the pid then we need to check the
socket name after the service restart.
- After restarting the radosgw service we need to wait few seconds
otherwise the socket won't be created.
- Update the wget parameters because the command is doing a loop.
We now use the same option than curl.
- The check_rest function doesn't test the radosgw at all due to
a wrong test command (test against a string) and always returns 0.
This needs to use the DOCKER_EXECS variable in order to execute the
command.
$ test 'wget http://192.168.100.11:8080'
$ echo $?
0
Also remove the test based on the ansible_fqdn because we only use
the ansible_hostname + rgw instance name.
Giulio Fidente [Wed, 19 Jun 2019 12:59:15 +0000 (14:59 +0200)]
Add radosgw_frontend_ssl_certificate parameter
This is necessary when configuring RGW with SSL because
in addition to passing specific frontend options, civetweb
appends the 's' character to the binding port and beast uses
ssl_endpoint instead of endpoint.
fmount [Wed, 19 Jun 2019 11:52:31 +0000 (13:52 +0200)]
Set grafana_server_addr fact for ipv6 scenarios.
As the bz1721914 describes, the grafana_server_addr
fact is not defined if ip_version used is ipv6.
This commit adds the ip_version condition to set
correctly this fact.
facts: fix bug in grafana_server_addr fact setting
If no grafana-server group is defined while an mgr group is, that task
will fail because `hostvars[groups[grafana_server_group_name][0]` can't
return anything since `groups['grafana-server']` will be a non existing
key.
To address this warning:
```
[DEPRECATION WARNING]: evaluating nfs_ganesha_dev as a bare variable, this
behaviour will go away and you might need to add |bool to the expression in the
future
```
Gabriel Ramirez [Tue, 25 Jun 2019 04:52:11 +0000 (21:52 -0700)]
validate.py: Fix alphabetical order on uca
Alphabetized ceph_repository_uca keys due to errors validating when
using UCA/queens repository on Ubuntu 16.04
An exception occurred during task execution. To see the full
traceback, use -vvv. The error was:
SchemaError: -> ceph_stable_repo_uca schema item is not
alphabetically ordered
Closes: #4154 Signed-off-by: Gabriel Ramirez <gabrielramirez1109@gmail.com>
This tries to first unmount any cephfs/nfs-ganesha mount point on client
nodes, then unmap any mapped rbd devices and finally it tries to remove
ceph kernel modules.
If it fails it means some resources are still busy and should be cleaned
manually before continuing to purge the cluster.
This is done early in the playbook so the cluster stays untouched until
everything is ready for that operation, otherwise if you try to redeploy
a cluster it could end up by getting confused by leftover from previous
deployment.
Dimitri Savineau [Thu, 20 Jun 2019 21:33:39 +0000 (17:33 -0400)]
ceph-handler: Fix OSD restart script
There's two big issues with the current OSD restart script.
1/ We try to test if the ceph osd daemon socket exists but we use a
wildcard for the socket name : /var/run/ceph/*.asok.
This fails because we usually have multiple ceph osd sockets (or
other ceph daemon collocated) present in /var/run/ceph directory.
Currently the test fails with:
bash: line xxx: [: too many arguments
But it doesn't stop the script execution.
Instead we can specify the full ceph osd socket name because we
already know the OSD id.
2/ The container filter pattern is wrong and could matches multiple
containers resulting the script to fail.
We use the filter with two different patterns. One is with the device
name (sda, sdb, ..) and the other one is with the OSD id (ceph-osd-0,
ceph-osd-15, ..).
In both case we could match more than needed.
Dimitri Savineau [Thu, 20 Jun 2019 18:04:24 +0000 (14:04 -0400)]
Change ansible_lsb by ansible_distribution_release
The ansible_lsb fact is based on the lsb package (lsb-base,
lsb-release or redhat-lsb-core).
If the package isn't installed on the remote host then the fact isn't
populated.
--------
"ansible_lsb": {},
--------
Switching to the ansible_distribution_release fact instead.
fpantano [Wed, 19 Jun 2019 17:13:37 +0000 (19:13 +0200)]
Add higher retry/delay defaults to check the quorum status.
As per bz1718981, this commit adds higher values to check
the quorum status. This is helpful for several OSP deployments
that fail during the scale up.
Dimitri Savineau [Thu, 20 Jun 2019 14:28:44 +0000 (10:28 -0400)]
ceph-volume: Set max open files limit on container
The ceph-volume lvm list command takes ages to complete when having
a lot of LV devices on containerized deployment.
For instance, with 25 OSDs on a node it takes 3 mins 44s to list the
OSD.
Adding the max open files limit to the container engine cli when
executing the ceph-volume command seems to improve a lot thee
execution time ~30s.
This was impacting the OSDs creation with ceph-volume (both filestore
and bluestore) when using multiple LV devices.
not sure exactly why since just before this task, mon1 seems to be well
UP otherwise it wouldn't have passed the task `waiting for the
containerized monitor to join the quorum`.
As a quick fix/workaround, let's add a retry which allows us to get
around this situation:
`parted_results` isn't used anymore in the playbook.
By the way, `parted` seems to cause issue because it changes the
ownership on devices:
```
root@osd0 ~]# ls -l /dev/sdc*
brw-rw----. 1 root disk 8, 32 Jun 11 08:53 /dev/sdc
brw-rw----. 1 ceph ceph 8, 33 Jun 11 08:53 /dev/sdc1
brw-rw----. 1 ceph ceph 8, 34 Jun 11 08:53 /dev/sdc2
[root@osd0 ~]# parted -s /dev/sdc print
Model: ATA QEMU HARDDISK (scsi)
Disk /dev/sdc: 53.7GB
Sector size (logical/physical): 512B/512B
Partition Table: gpt
Disk Flags:
Number Start End Size File system Name Flags
1 1049kB 1075MB 1074MB ceph block.db
2 1075MB 2149MB 1074MB ceph block.db
[root@osd0 ~]# #We can see ownerships have changed from ceph:ceph to root:disk:
[root@osd0 ~]# ls -l /dev/sdc*
brw-rw----. 1 root disk 8, 32 Jun 11 08:57 /dev/sdc
brw-rw----. 1 root disk 8, 33 Jun 11 08:57 /dev/sdc1
brw-rw----. 1 root disk 8, 34 Jun 11 08:57 /dev/sdc2
[root@osd0 ~]#
```
rolling_update: only mask and stop unit in mgr part
Otherwise it fails like following:
```
fatal: [mon0]: FAILED! => changed=false
msg: |-
Unable to enable service ceph-mgr@mon0: Failed to execute operation: Cannot send after transport endpoint shutdown
```
Dimitri Savineau [Fri, 14 Jun 2019 21:31:39 +0000 (17:31 -0400)]
tests: Update ansible ssh_args variable
Because we're using vagrant, a ssh config file will be created for
each nodes with options like user, host, port, identity, etc...
But via tox we're override ANSIBLE_SSH_ARGS to use this file. This
remove the default value set in ansible.cfg.
Also adding PreferredAuthentications=publickey because CentOS/RHEL
servers are configured with GSSAPIAuthenticationis enabled for ssh
server forcing the client to make a PTR DNS query.
projects / ceph-ansible.git / log