summaryrefslogtreecommitdiffstats
path: root/roles
diff options
context:
space:
mode:
Diffstat (limited to 'roles')
-rw-r--r--roles/docker/handlers/main.yml5
-rw-r--r--roles/openshift_health_checker/action_plugins/openshift_health_check.py5
-rw-r--r--roles/openshift_health_checker/openshift_checks/disk_availability.py120
-rw-r--r--roles/openshift_health_checker/test/action_plugin_test.py12
-rw-r--r--roles/openshift_health_checker/test/disk_availability_test.py27
-rw-r--r--roles/openshift_logging_elasticsearch/tasks/main.yaml10
-rw-r--r--roles/openshift_logging_kibana/tasks/main.yaml5
7 files changed, 120 insertions, 64 deletions
diff --git a/roles/docker/handlers/main.yml b/roles/docker/handlers/main.yml
index 7f91afb37..3a4f4ba92 100644
--- a/roles/docker/handlers/main.yml
+++ b/roles/docker/handlers/main.yml
@@ -4,6 +4,11 @@
systemd:
name: "{{ openshift.docker.service_name }}"
state: restarted
+ register: r_docker_restart_docker_result
+ until: not r_docker_restart_docker_result | failed
+ retries: 1
+ delay: 30
+
when: not docker_service_status_changed | default(false) | bool
- name: restart udev
diff --git a/roles/openshift_health_checker/action_plugins/openshift_health_check.py b/roles/openshift_health_checker/action_plugins/openshift_health_check.py
index a62e4331e..0390dc82e 100644
--- a/roles/openshift_health_checker/action_plugins/openshift_health_check.py
+++ b/roles/openshift_health_checker/action_plugins/openshift_health_check.py
@@ -38,14 +38,13 @@ class ActionModule(ActionBase):
try:
known_checks = self.load_known_checks()
+ args = self._task.args
+ resolved_checks = resolve_checks(args.get("checks", []), known_checks.values())
except OpenShiftCheckException as e:
result["failed"] = True
result["msg"] = str(e)
return result
- args = self._task.args
- resolved_checks = resolve_checks(args.get("checks", []), known_checks.values())
-
result["checks"] = check_results = {}
user_disabled_checks = [
diff --git a/roles/openshift_health_checker/openshift_checks/disk_availability.py b/roles/openshift_health_checker/openshift_checks/disk_availability.py
index 962148cb8..e93e81efa 100644
--- a/roles/openshift_health_checker/openshift_checks/disk_availability.py
+++ b/roles/openshift_health_checker/openshift_checks/disk_availability.py
@@ -1,9 +1,12 @@
-# pylint: disable=missing-docstring
+"""Check that there is enough disk space in predefined paths."""
+
+import os.path
+import tempfile
+
from openshift_checks import OpenShiftCheck, OpenShiftCheckException, get_var
-from openshift_checks.mixins import NotContainerizedMixin
-class DiskAvailability(NotContainerizedMixin, OpenShiftCheck):
+class DiskAvailability(OpenShiftCheck):
"""Check that recommended disk space is available before a first-time install."""
name = "disk_availability"
@@ -12,56 +15,101 @@ class DiskAvailability(NotContainerizedMixin, OpenShiftCheck):
# Values taken from the official installation documentation:
# https://docs.openshift.org/latest/install_config/install/prerequisites.html#system-requirements
recommended_disk_space_bytes = {
- "masters": 40 * 10**9,
- "nodes": 15 * 10**9,
- "etcd": 20 * 10**9,
+ '/var': {
+ 'masters': 40 * 10**9,
+ 'nodes': 15 * 10**9,
+ 'etcd': 20 * 10**9,
+ },
+ # Used to copy client binaries into,
+ # see roles/openshift_cli/library/openshift_container_binary_sync.py.
+ '/usr/local/bin': {
+ 'masters': 1 * 10**9,
+ 'nodes': 1 * 10**9,
+ 'etcd': 1 * 10**9,
+ },
+ # Used as temporary storage in several cases.
+ tempfile.gettempdir(): {
+ 'masters': 1 * 10**9,
+ 'nodes': 1 * 10**9,
+ 'etcd': 1 * 10**9,
+ },
}
@classmethod
def is_active(cls, task_vars):
"""Skip hosts that do not have recommended disk space requirements."""
group_names = get_var(task_vars, "group_names", default=[])
- has_disk_space_recommendation = bool(set(group_names).intersection(cls.recommended_disk_space_bytes))
+ active_groups = set()
+ for recommendation in cls.recommended_disk_space_bytes.values():
+ active_groups.update(recommendation.keys())
+ has_disk_space_recommendation = bool(active_groups.intersection(group_names))
return super(DiskAvailability, cls).is_active(task_vars) and has_disk_space_recommendation
def run(self, tmp, task_vars):
group_names = get_var(task_vars, "group_names")
ansible_mounts = get_var(task_vars, "ansible_mounts")
- free_bytes = self.openshift_available_disk(ansible_mounts)
-
- recommended_min = max(self.recommended_disk_space_bytes.get(name, 0) for name in group_names)
- configured_min = int(get_var(task_vars, "openshift_check_min_host_disk_gb", default=0)) * 10**9
- min_free_bytes = configured_min or recommended_min
-
- if free_bytes < min_free_bytes:
- return {
- 'failed': True,
- 'msg': (
- 'Available disk space ({:.1f} GB) for the volume containing '
- '"/var" is below minimum recommended space ({:.1f} GB)'
- ).format(float(free_bytes) / 10**9, float(min_free_bytes) / 10**9)
+ ansible_mounts = {mount['mount']: mount for mount in ansible_mounts}
+
+ user_config = get_var(task_vars, "openshift_check_min_host_disk_gb", default={})
+ try:
+ # For backwards-compatibility, if openshift_check_min_host_disk_gb
+ # is a number, then it overrides the required config for '/var'.
+ number = float(user_config)
+ user_config = {
+ '/var': {
+ 'masters': number,
+ 'nodes': number,
+ 'etcd': number,
+ },
}
+ except TypeError:
+ # If it is not a number, then it should be a nested dict.
+ pass
+
+ # TODO: as suggested in
+ # https://github.com/openshift/openshift-ansible/pull/4436#discussion_r122180021,
+ # maybe we could support checking disk availability in paths that are
+ # not part of the official recommendation but present in the user
+ # configuration.
+ for path, recommendation in self.recommended_disk_space_bytes.items():
+ free_bytes = self.free_bytes(path, ansible_mounts)
+ recommended_bytes = max(recommendation.get(name, 0) for name in group_names)
+
+ config = user_config.get(path, {})
+ # NOTE: the user config is in GB, but we compare bytes, thus the
+ # conversion.
+ config_bytes = max(config.get(name, 0) for name in group_names) * 10**9
+ recommended_bytes = config_bytes or recommended_bytes
+
+ if free_bytes < recommended_bytes:
+ free_gb = float(free_bytes) / 10**9
+ recommended_gb = float(recommended_bytes) / 10**9
+ return {
+ 'failed': True,
+ 'msg': (
+ 'Available disk space in "{}" ({:.1f} GB) '
+ 'is below minimum recommended ({:.1f} GB)'
+ ).format(path, free_gb, recommended_gb)
+ }
return {}
@staticmethod
- def openshift_available_disk(ansible_mounts):
- """Determine the available disk space for an OpenShift installation.
-
- ansible_mounts should be a list of dicts like the 'setup' Ansible module
- returns.
- """
- # priority list in descending order
- supported_mnt_paths = ["/var", "/"]
- available_mnts = {mnt.get("mount"): mnt for mnt in ansible_mounts}
+ def free_bytes(path, ansible_mounts):
+ """Return the size available in path based on ansible_mounts."""
+ mount_point = path
+ # arbitry value to prevent an infinite loop, in the unlike case that '/'
+ # is not in ansible_mounts.
+ max_depth = 32
+ while mount_point not in ansible_mounts and max_depth > 0:
+ mount_point = os.path.dirname(mount_point)
+ max_depth -= 1
try:
- for path in supported_mnt_paths:
- if path in available_mnts:
- return available_mnts[path]["size_available"]
+ free_bytes = ansible_mounts[mount_point]['size_available']
except KeyError:
- pass
+ known_mounts = ', '.join('"{}"'.format(mount) for mount in sorted(ansible_mounts)) or 'none'
+ msg = 'Unable to determine disk availability for "{}". Known mount points: {}.'
+ raise OpenShiftCheckException(msg.format(path, known_mounts))
- paths = ''.join(sorted(available_mnts)) or 'none'
- msg = "Unable to determine available disk space. Paths mounted: {}.".format(paths)
- raise OpenShiftCheckException(msg)
+ return free_bytes
diff --git a/roles/openshift_health_checker/test/action_plugin_test.py b/roles/openshift_health_checker/test/action_plugin_test.py
index 6ebf0ebb2..9383b233c 100644
--- a/roles/openshift_health_checker/test/action_plugin_test.py
+++ b/roles/openshift_health_checker/test/action_plugin_test.py
@@ -59,7 +59,7 @@ def failed(result, msg_has=None):
if msg_has is not None:
assert 'msg' in result
for term in msg_has:
- assert term in result['msg']
+ assert term.lower() in result['msg'].lower()
return result.get('failed', False)
@@ -178,6 +178,16 @@ def test_action_plugin_run_check_exception(plugin, task_vars, monkeypatch):
assert not skipped(result)
+def test_action_plugin_resolve_checks_exception(plugin, task_vars, monkeypatch):
+ monkeypatch.setattr(plugin, 'load_known_checks', lambda: {})
+
+ result = plugin.run(tmp=None, task_vars=task_vars)
+
+ assert failed(result, msg_has=['unknown', 'name'])
+ assert not changed(result)
+ assert not skipped(result)
+
+
@pytest.mark.parametrize('names,all_checks,expected', [
([], [], set()),
(
diff --git a/roles/openshift_health_checker/test/disk_availability_test.py b/roles/openshift_health_checker/test/disk_availability_test.py
index b353fa610..945b9eafc 100644
--- a/roles/openshift_health_checker/test/disk_availability_test.py
+++ b/roles/openshift_health_checker/test/disk_availability_test.py
@@ -3,22 +3,19 @@ import pytest
from openshift_checks.disk_availability import DiskAvailability, OpenShiftCheckException
-@pytest.mark.parametrize('group_names,is_containerized,is_active', [
- (['masters'], False, True),
- # ensure check is skipped on containerized installs
- (['masters'], True, False),
- (['nodes'], False, True),
- (['etcd'], False, True),
- (['masters', 'nodes'], False, True),
- (['masters', 'etcd'], False, True),
- ([], False, False),
- (['lb'], False, False),
- (['nfs'], False, False),
+@pytest.mark.parametrize('group_names,is_active', [
+ (['masters'], True),
+ (['nodes'], True),
+ (['etcd'], True),
+ (['masters', 'nodes'], True),
+ (['masters', 'etcd'], True),
+ ([], False),
+ (['lb'], False),
+ (['nfs'], False),
])
-def test_is_active(group_names, is_containerized, is_active):
+def test_is_active(group_names, is_active):
task_vars = dict(
group_names=group_names,
- openshift=dict(common=dict(is_containerized=is_containerized)),
)
assert DiskAvailability.is_active(task_vars=task_vars) == is_active
@@ -38,7 +35,7 @@ def test_cannot_determine_available_disk(ansible_mounts, extra_words):
with pytest.raises(OpenShiftCheckException) as excinfo:
check.run(tmp=None, task_vars=task_vars)
- for word in 'determine available disk'.split() + extra_words:
+ for word in 'determine disk availability'.split() + extra_words:
assert word in str(excinfo.value)
@@ -81,7 +78,7 @@ def test_cannot_determine_available_disk(ansible_mounts, extra_words):
[{
# not enough space on / ...
'mount': '/',
- 'size_available': 0,
+ 'size_available': 2 * 10**9,
}, {
# ... but enough on /var
'mount': '/var',
diff --git a/roles/openshift_logging_elasticsearch/tasks/main.yaml b/roles/openshift_logging_elasticsearch/tasks/main.yaml
index d9ac52cb7..68726aa78 100644
--- a/roles/openshift_logging_elasticsearch/tasks/main.yaml
+++ b/roles/openshift_logging_elasticsearch/tasks/main.yaml
@@ -173,9 +173,8 @@
selector:
component: "{{ es_component }}"
provider: openshift
- # pending #4091
- #labels:
- #- logging-infra: 'support'
+ labels:
+ logging-infra: 'support'
ports:
- port: 9300
@@ -187,9 +186,8 @@
selector:
component: "{{ es_component }}"
provider: openshift
- # pending #4091
- #labels:
- #- logging-infra: 'support'
+ labels:
+ logging-infra: 'support'
ports:
- port: 9200
targetPort: "restapi"
diff --git a/roles/openshift_logging_kibana/tasks/main.yaml b/roles/openshift_logging_kibana/tasks/main.yaml
index bae55ffaa..93cb82793 100644
--- a/roles/openshift_logging_kibana/tasks/main.yaml
+++ b/roles/openshift_logging_kibana/tasks/main.yaml
@@ -89,9 +89,8 @@
selector:
component: "{{ kibana_component }}"
provider: openshift
- # pending #4091
- #labels:
- #- logging-infra: 'support'
+ labels:
+ logging-infra: 'support'
ports:
- port: 443
targetPort: "oaproxy"