mgr/cephadm: allow having prometheus bind on specific network

To avoid port conflicts in certain setups where
something else is using the same port on a different
network on the host.

Signed-off-by: Adam King <adking@redhat.com>
(cherry picked from commit e6943dd61374ee8ce337751e9ef9ba55f10a83ed)

diff --git a/src/cephadm/cephadmlib/daemons/monitoring.py b/src/cephadm/cephadmlib/daemons/monitoring.py
index e0667853dd7b5732c02d718458ef2e98d5924949..e6ad2412a7cbaf255d0308e594dacc70aa4758b0 100644 (file)
--- a/src/cephadm/cephadmlib/daemons/monitoring.py
+++ b/src/cephadm/cephadmlib/daemons/monitoring.py
@@ -251,9 +251,11 @@ class Monitoring(ContainerDaemonForm):
                     ip = meta['ip']
                 if 'ports' in meta and meta['ports']:
                     port = meta['ports'][0]
-            r += [f'--web.listen-address={ip}:{port}']
             if daemon_type == 'prometheus':
                 config = fetch_configs(ctx)
+                ip_to_bind_to = config.get('ip_to_bind_to', '')
+                if ip_to_bind_to:
+                    ip = ip_to_bind_to
                 retention_time = config.get('retention_time', '15d')
                 retention_size = config.get(
                     'retention_size', '0'
@@ -270,6 +272,7 @@ class Monitoring(ContainerDaemonForm):
                     addr = next(iter(ipv4_addrs or ipv6_addrs), None)
                     host = wrap_ipv6(addr) if addr else host
                 r += [f'--web.external-url={scheme}://{host}:{port}']
+            r += [f'--web.listen-address={ip}:{port}']
         if daemon_type == 'alertmanager':
             config = fetch_configs(ctx)
             peers = config.get('peers', list())  # type: ignore

diff --git a/src/cephadm/tests/test_deploy.py b/src/cephadm/tests/test_deploy.py
index cfde3fbce0aab80210f3fc8cd153dbf012edf55e..9a6da3aad4b958601f4f93f6ff52028b40bc6bc2 100644 (file)
--- a/src/cephadm/tests/test_deploy.py
+++ b/src/cephadm/tests/test_deploy.py
@@ -305,6 +305,7 @@ def test_deploy_a_monitoring_container(cephadm_fs, funkypatch):
             'files': {
                 'prometheus.yml': 'bettercallherc',
             },
+            'ip_to_bind_to': '1.2.3.4'
         }
         _cephadm._common_deploy(ctx)
 
@@ -314,7 +315,7 @@ def test_deploy_a_monitoring_container(cephadm_fs, funkypatch):
         runfile_lines = f.read().splitlines()
     assert 'podman' in runfile_lines[-1]
     assert runfile_lines[-1].endswith(
-        'quay.io/titans/prometheus:latest --config.file=/etc/prometheus/prometheus.yml --storage.tsdb.path=/prometheus --web.listen-address=:9095 --storage.tsdb.retention.time=15d --storage.tsdb.retention.size=0 --web.external-url=http://10.10.10.10:9095'
+        'quay.io/titans/prometheus:latest --config.file=/etc/prometheus/prometheus.yml --storage.tsdb.path=/prometheus --storage.tsdb.retention.time=15d --storage.tsdb.retention.size=0 --web.external-url=http://10.10.10.10:9095 --web.listen-address=1.2.3.4:9095'
     )
     assert '--user 8765' in runfile_lines[-1]
     assert f'-v /var/lib/ceph/{fsid}/prometheus.fire/etc/prometheus:/etc/prometheus:Z' in runfile_lines[-1]

diff --git a/src/pybind/mgr/cephadm/services/monitoring.py b/src/pybind/mgr/cephadm/services/monitoring.py
index 9aa503b5bcefa26abe0fca6fafc511d05a082af9..cc8ebd1d628ed27453b33fdc361100fece95e787 100644 (file)
--- a/src/pybind/mgr/cephadm/services/monitoring.py
+++ b/src/pybind/mgr/cephadm/services/monitoring.py
@@ -431,6 +431,13 @@ class PrometheusService(CephadmService):
             'ceph_exporter_sd_url': ceph_exporter_sd_url
         }
 
+        ip_to_bind_to = ''
+        if spec.only_bind_port_on_networks and spec.networks:
+            assert daemon_spec.host is not None
+            ip_to_bind_to = self.mgr.get_first_matching_network_ip(daemon_spec.host, spec) or ''
+            if ip_to_bind_to:
+                daemon_spec.port_ips = {str(port): ip_to_bind_to}
+
         web_context = {
             'prometheus_web_user': prometheus_user,
             'prometheus_web_password': password_hash(prometheus_password),
@@ -457,6 +464,7 @@ class PrometheusService(CephadmService):
                     },
                     'retention_time': retention_time,
                     'retention_size': retention_size,
+                    'ip_to_bind_to': ip_to_bind_to,
                     'web_config': '/etc/prometheus/web.yml'
                 }
         else:
@@ -465,7 +473,8 @@ class PrometheusService(CephadmService):
                     'prometheus.yml': self.mgr.template.render('services/prometheus/prometheus.yml.j2', context)
                 },
                 'retention_time': retention_time,
-                'retention_size': retention_size
+                'retention_size': retention_size,
+                'ip_to_bind_to': ip_to_bind_to
             }
 
         # include alerts, if present in the container

diff --git a/src/pybind/mgr/cephadm/tests/test_services.py b/src/pybind/mgr/cephadm/tests/test_services.py
index dbab022058d424e132d8a8e26f0de1198f129884..73e91353384535f840a3546f5a0346e952a3e517 100644 (file)
--- a/src/pybind/mgr/cephadm/tests/test_services.py
+++ b/src/pybind/mgr/cephadm/tests/test_services.py
@@ -682,7 +682,9 @@ class TestMonitoring:
                                                              keepalived_password='12345',
                                                              virtual_ip="1.2.3.4/32",
                                                              backend_service='rgw.foo')) as _, \
-                    with_service(cephadm_module, PrometheusSpec('prometheus')) as _:
+                    with_service(cephadm_module, PrometheusSpec('prometheus',
+                                                                networks=['1.2.3.0/24'],
+                                                                only_bind_port_on_networks=True)) as _:
 
                 y = dedent("""
                 # This file is generated by cephadm.
@@ -730,11 +732,12 @@ class TestMonitoring:
                         "deploy_arguments": [],
                         "params": {
                             'tcp_ports': [9095],
+                            'port_ips': {'8765': '1.2.3.1'}
                         },
                         "meta": {
                             'service_name': 'prometheus',
                             'ports': [9095],
-                            'ip': None,
+                            'ip': '1.2.3.1',
                             'deployed_by': [],
                             'rank': None,
                             'rank_generation': None,
@@ -748,6 +751,7 @@ class TestMonitoring:
                             },
                             'retention_time': '15d',
                             'retention_size': '0',
+                            'ip_to_bind_to': '1.2.3.1',
                         },
                     }),
                 )
@@ -909,6 +913,7 @@ class TestMonitoring:
                             },
                             'retention_time': '15d',
                             'retention_size': '0',
+                            'ip_to_bind_to': '',
                             'web_config': '/etc/prometheus/web.yml',
                         },
                     }),

diff --git a/src/python-common/ceph/deployment/service_spec.py b/src/python-common/ceph/deployment/service_spec.py
index 6808567500f81360524cb074eb817961230d8975..d5a59c9b12158bc78a83e34cd144f7a5b6f0e2b9 100644 (file)
--- a/src/python-common/ceph/deployment/service_spec.py
+++ b/src/python-common/ceph/deployment/service_spec.py
@@ -1885,6 +1885,7 @@ class PrometheusSpec(MonitoringSpec):
                  preview_only: bool = False,
                  config: Optional[Dict[str, str]] = None,
                  networks: Optional[List[str]] = None,
+                 only_bind_port_on_networks: bool = False,
                  port: Optional[int] = None,
                  retention_time: Optional[str] = None,
                  retention_size: Optional[str] = None,
@@ -1902,6 +1903,7 @@ class PrometheusSpec(MonitoringSpec):
 
         self.retention_time = retention_time.strip() if retention_time else None
         self.retention_size = retention_size.strip() if retention_size else None
+        self.only_bind_port_on_networks = only_bind_port_on_networks
 
     def validate(self) -> None:
         super(PrometheusSpec, self).validate()