Merge lp:~james-page/charms/precise/quantum-gateway/ha-fixes into lp:~openstack-charmers/charms/precise/quantum-gateway/ha-support

[James Page]

Quantum has undergone some refactoring upstream and now includes a scheduler which spread dhcp and l3 agent services across all gateway nodes.

This invalidates the original HA approach; HA was on the plan for the scheduler - trying to find out if it will make grizzly.

Other than that fixed up a few other minor issues and added support for HTTPS on the quantum API service itself.

[Adam Gandelman (gandelman-a)]

James-

This looks good to me, but I can't seem to actually push the merge back?

$ bzr push :parent
bzr: ERROR: Cannot lock LockDir(chroot-93200144:///~openstack-charmers/charms/precise/quantum-gateway/ha-support/.bzr/branch/lock): Transport operation not possible: readonly transport

=== modified file 'config.yaml'
--- config.yaml	2013-01-18 10:09:25 +0000
+++ config.yaml	2013-03-11 14:16:22 +0000
@@ -7,6 +7,7 @@
       Supported values include:
       .
         ovs - OpenVSwitch
+        nvp - Nicira Network Virtualization Platform
   ext-port:
     type: string
     description: |

=== removed symlink 'hooks/cluster-relation-changed'
=== target was u'hooks.py'
=== modified file 'hooks/hooks.py'
--- hooks/hooks.py	2013-02-18 17:50:16 +0000
+++ hooks/hooks.py	2013-03-11 14:16:22 +0000
@@ -24,10 +24,11 @@
 def config_changed():
     if PLUGIN in qutils.GATEWAY_PKGS.keys():
         render_quantum_conf()
-        render_plugin_conf()
+        render_dhcp_agent_conf()
         render_l3_agent_conf()
         render_metadata_agent_conf()
         render_metadata_api_conf()
+        render_plugin_conf()
         if PLUGIN == qutils.OVS:
             qutils.add_bridge(qutils.INT_BRIDGE)
             qutils.add_bridge(qutils.EXT_BRIDGE)
@@ -58,6 +59,16 @@
                        )
 
 
+def render_dhcp_agent_conf():
+    if (os.path.exists(qutils.DHCP_AGENT_CONF)):
+        with open(qutils.DHCP_AGENT_CONF, "w") as conf:
+            conf.write(utils.render_template(
+                            os.path.basename(qutils.DHCP_AGENT_CONF),
+                            {}
+                            )
+                       )
+
+
 def render_metadata_agent_conf():
     context = get_keystone_conf()
     if (context and
@@ -136,6 +147,8 @@
                                                    unit, relid),
                 "quantum_port": utils.relation_get('quantum_port',
                                                    unit, relid),
+                "quantum_url": utils.relation_get('quantum_url',
+                                                   unit, relid),
                 "region": utils.relation_get('region',
                                              unit, relid)
                 }
@@ -197,6 +210,7 @@
 
 
 def amqp_changed():
+    render_dhcp_agent_conf()
     render_quantum_conf()
     render_metadata_api_conf()
     restart_agents()
@@ -222,62 +236,31 @@
 
 
 def nm_changed():
+    render_dhcp_agent_conf()
     render_l3_agent_conf()
     render_metadata_agent_conf()
     render_metadata_api_conf()
+    store_ca_cert()
     restart_agents()
 
 
+def store_ca_cert():
+    ca_cert = get_ca_cert()
+    if ca_cert:
+        utils.install_ca(ca_cert)
+
+
+def get_ca_cert():
+    for relid in utils.relation_ids('quantum-network-service'):
+        for unit in utils.relation_list(relid):
+            ca_cert = utils.relation_get('ca_cert', unit, relid)
+            if ca_cert:
+                return ca_cert
+    return None
+
+
 def restart_agents():
-    if utils.eligible_leader(RESOURCE_GROUP):
-        utils.restart(*qutils.GATEWAY_AGENTS[PLUGIN])
-    else:
-        # Stop any clustered agents and flush local config
-        # from network namespaces, ovs ports and dnsmasq instances
-        utils.stop(*qutils.CLUSTERED_AGENTS[PLUGIN])
-        utils.restart(*qutils.STANDALONE_AGENTS[PLUGIN])
-        qutils.flush_local_configuration()
-
-
-RESOURCE_GROUP = 'grp_quantum_services'
-
-
-def ha_relation_joined():
-    # init services that will be clusterized. Used to disable init scripts
-    # Used when resources have upstart jobs that are needed to be disabled.
-    # resource_name:init_script_name
-    init_services = {'res_quantum_dhcp_agent': 'quantum-dhcp-agent',
-                     'res_quantum_l3_agent': 'quantum-l3-agent'}
-
-    # Obtain resources
-    # TODO: Just use upstart for the time being
-    #resources = {'res_quantum_dhcp_agent': 'ocf:openstack:quantum-agent-dhcp',
-    #             'res_quantum_l3_agent': 'ocf:openstack:quantum-agent-l3'}
-    resources = {'res_quantum_dhcp_agent': 'upstart:quantum-dhcp-agent',
-                 'res_quantum_l3_agent': 'upstart:quantum-l3-agent'}
-    # TODO: monitors are currently disabled as this creates issues
-    #       when forming the cluster.
-    resource_params = {'res_quantum_dhcp_agent':
-                            'op monitor interval="5s"',
-                       'res_quantum_l3_agent':
-                            'op monitor interval="5s"'}
-    groups = {
-        RESOURCE_GROUP:
-            'res_quantum_dhcp_agent res_quantum_l3_agent'
-        }
-    # set relation values
-    utils.relation_set(resources=resources,
-                       resource_params=resource_params,
-                       init_services=init_services,
-                       groups=groups,
-                       corosync_bindiface=utils.config_get('ha-bindiface'),
-                       corosync_mcastport=utils.config_get('ha-mcastport'))
-
-
-def cluster_changed():
-    if not utils.eligible_leader(RESOURCE_GROUP):
-        utils.stop(*qutils.CLUSTERED_AGENTS[PLUGIN])
-        qutils.flush_local_configuration()
+    utils.restart(*qutils.GATEWAY_AGENTS[PLUGIN])
 
 
 utils.do_hooks({
@@ -288,9 +271,7 @@
     "shared-db-relation-changed": db_changed,
     "amqp-relation-joined": amqp_joined,
     "amqp-relation-changed": amqp_changed,
-    "quantum-network-service-relation-changed": nm_changed,
-    "ha-relation-joined": ha_relation_joined,
-    "cluster-relation-changed": cluster_changed
+    "quantum-network-service-relation-changed": nm_changed
     })
 
 sys.exit(0)

=== modified file 'hooks/quantum_utils.py'
--- hooks/quantum_utils.py	2013-02-18 09:14:49 +0000
+++ hooks/quantum_utils.py	2013-03-11 14:16:22 +0000
@@ -6,24 +6,17 @@
 
 
 OVS = "ovs"
-NVP = "nvp"
 
 OVS_PLUGIN = \
     "quantum.plugins.openvswitch.ovs_quantum_plugin.OVSQuantumPluginV2"
-NVP_PLUGIN = \
-    "quantum.plugins.nicira.nicira_nvp_plugin.QuantumPlugin.NvpPluginV2"
 CORE_PLUGIN = {
     OVS: OVS_PLUGIN,
-    NVP: NVP_PLUGIN
     }
 
 OVS_PLUGIN_CONF = \
     "/etc/quantum/plugins/openvswitch/ovs_quantum_plugin.ini"
-NVP_PLUGIN_CONF = \
-    "/etc/quantum/plugins/nicira/nvp.ini"
 PLUGIN_CONF = {
     OVS: OVS_PLUGIN_CONF,
-    NVP: NVP_PLUGIN_CONF
     }
 
 GATEWAY_PKGS = {
@@ -34,19 +27,8 @@
         'python-mysqldb',
         "nova-api-metadata"
         ],
-    NVP: [
-        "quantum-plugin-nicira",
-        "quantum-l3-agent",
-        "quantum-dhcp-agent",
-        'python-mysqldb',
-        "nova-api-metadata"
-        ]
     }
 
-# TODO: conditionally add quantum-metadata-agent if
-# running 2013.1 onwards. OR add some overrides
-# start on starting quantum-l3-agent
-# stop on stopping quantum-l3-agent
 GATEWAY_AGENTS = {
     OVS: [
         "quantum-plugin-openvswitch-agent",
@@ -54,33 +36,8 @@
         "quantum-dhcp-agent",
         "nova-api-metadata"
         ],
-    NVP: [
-        "quantum-l3-agent",
-        "quantum-dhcp-agent",
-        "nova-api-metadata"
-        ]
-    }
-
-CLUSTERED_AGENTS = {
-    OVS: [
-        "quantum-l3-agent",
-        "quantum-dhcp-agent",
-        ],
-    NVP: [
-        "quantum-l3-agent",
-        "quantum-dhcp-agent",
-        ]
-    }
-
-STANDALONE_AGENTS = {
-    OVS: [
-        "quantum-plugin-openvswitch-agent",
-        "nova-api-metadata"
-        ],
-    NVP: [
-        "nova-api-metadata"
-        ]
-    }
+    }
+
 
 if get_os_version('quantum-common') >= "2013.1":
     for plugin in GATEWAY_AGENTS:

=== modified file 'hooks/utils.py'
--- hooks/utils.py	2013-02-08 17:51:20 +0000
+++ hooks/utils.py	2013-03-11 14:16:22 +0000
@@ -12,6 +12,7 @@
 import socket
 import sys
 import apt_pkg as apt
+import base64
 
 
 def do_hooks(hooks):
@@ -86,8 +87,8 @@
     if source.startswith('cloud:'):
         install('ubuntu-cloud-keyring')
         pocket = source.split(':')[1]
-        with open('/etc/apt/sources.list.d/cloud-archive.list', 'w') as apt:
-            apt.write(CLOUD_ARCHIVE.format(CLOUD_ARCHIVE_POCKETS[pocket]))
+        with open('/etc/apt/sources.list.d/cloud-archive.list', 'w') as sfile:
+            sfile.write(CLOUD_ARCHIVE.format(CLOUD_ARCHIVE_POCKETS[pocket]))
     if source.startswith('deb'):
         l = len(source.split('|'))
         if l == 2:
@@ -101,8 +102,8 @@
         elif l == 1:
             apt_line = source
 
-        with open('/etc/apt/sources.list.d/quantum.list', 'w') as apt:
-            apt.write(apt_line + "\n")
+        with open('/etc/apt/sources.list.d/quantum.list', 'w') as sfile:
+            sfile.write(apt_line + "\n")
     cmd = [
         'apt-get',
         'update'
@@ -214,38 +215,15 @@
         return hostname
     except socket.error:
         pass
-    try:
-        answers = dns.resolver.query(hostname, 'A')
-        if answers:
-            return answers[0].address
-    except dns.resolver.NXDOMAIN:
-        pass
-    return None
-
-
-CLUSTER_RESOURCES = {
-    'quantum-dhcp-agent': 'res_quantum_dhcp_agent',
-    'quantum-l3-agent': 'res_quantum_l3_agent'
-    }
-
-HAMARKER = '/var/lib/juju/haconfigured'
+    answers = dns.resolver.query(hostname, 'A')
+    if answers:
+        return answers[0].address
+    else:
+        return None
 
 
 def _service_ctl(service, action):
-    if (os.path.exists(HAMARKER) and
-        os.path.exists(os.path.join('/etc/init/',
-                                   '{}.override'.format(service))) and
-        service in CLUSTER_RESOURCES):
-        hostname = str(subprocess.check_output(['hostname'])).strip()
-        service_status = \
-            subprocess.check_output(['crm', 'resource', 'show',
-                                     CLUSTER_RESOURCES[service]])
-        # Only restart if we are the node that owns the service
-        if hostname in service_status:
-            subprocess.check_call(['crm', 'resource', action,
-                                  CLUSTER_RESOURCES[service]])
-    else:
-        subprocess.call(['service', service, action])
+    subprocess.call(['service', service, action])
 
 
 def restart(*services):
@@ -273,55 +251,8 @@
         return None
 
 
-def is_clustered():
-    for r_id in (relation_ids('ha') or []):
-        for unit in (relation_list(r_id) or []):
-            clustered = relation_get('clustered',
-                                     unit=unit,
-                                     rid=r_id)
-            if clustered:
-                return True
-    return False
-
-
-def is_leader(resource):
-    if os.path.exists('/usr/sbin/crm'):
-        try:
-            status = subprocess.check_output(['crm', 'resource',
-                                              'show', resource])
-            hostname = get_unit_hostname()
-            if hostname in status:
-                return True
-        except subprocess.CalledProcessError:
-            pass
-    return False
-
-
-def peer_units():
-    peers = []
-    for r_id in (relation_ids('cluster') or []):
-        for unit in (relation_list(r_id) or []):
-            peers.append(unit)
-    return peers
-
-
-def oldest_peer(peers):
-    local_unit_no = int(os.getenv('JUJU_UNIT_NAME').split('/')[1])
-    for peer in peers:
-        remote_unit_no = int(peer.split('/')[1])
-        if remote_unit_no < local_unit_no:
-            return False
-    return True
-
-
-def eligible_leader(resource):
-    if is_clustered():
-        if not is_leader(resource):
-            juju_log('INFO', 'Deferring action to CRM leader.')
-            return False
-    else:
-        peers = peer_units()
-        if peers and not oldest_peer(peers):
-            juju_log('INFO', 'Deferring action to oldest service unit.')
-            return False
-    return True
+def install_ca(ca_cert):
+    with open('/usr/local/share/ca-certificates/keystone_juju_ca_cert.crt',
+              'w') as crt:
+        crt.write(base64.b64decode(ca_cert))
+    subprocess.check_call(['update-ca-certificates', '--fresh'])

=== modified file 'metadata.yaml'
--- metadata.yaml	2013-02-08 16:10:48 +0000
+++ metadata.yaml	2013-03-11 14:16:22 +0000
@@ -20,10 +20,4 @@
   shared-db:
     interface: mysql-shared
   amqp:
-    interface: rabbitmq
-  ha:
-    interface: hacluster
-    scope: container
-peers:
-  cluster:
-    interface: quantum-gateway-ha
+    interface: rabbitmq
\ No newline at end of file

=== modified file 'templates/nova.conf'
--- templates/nova.conf	2013-01-18 12:39:10 +0000
+++ templates/nova.conf	2013-03-11 14:16:22 +0000
@@ -18,7 +18,7 @@
 # Access to quantum API services
 network_api_class=nova.network.quantumv2.api.API
 quantum_auth_strategy=keystone
-quantum_url=http://{{ quantum_host }}:{{ quantum_port }}
+quantum_url={{ quantum_url }}
 quantum_admin_tenant_name={{ service_tenant }}
 quantum_admin_username={{ service_username }}
 quantum_admin_password={{ service_password }}

=== modified file 'templates/ovs_quantum_plugin.ini'
--- templates/ovs_quantum_plugin.ini	2012-11-05 11:59:27 +0000
+++ templates/ovs_quantum_plugin.ini	2013-03-11 14:16:22 +0000
@@ -7,5 +7,5 @@
 enable_tunneling = True
 tunnel_id_ranges = 1:1000
 [AGENT]
-polling_interval = 2
+polling_interval = 10
 root_helper = sudo /usr/bin/quantum-rootwrap /etc/quantum/rootwrap.conf

=== modified file 'templates/quantum.conf'
--- templates/quantum.conf	2013-02-18 09:14:49 +0000
+++ templates/quantum.conf	2013-03-11 14:16:22 +0000
@@ -14,7 +14,7 @@
 list_notifier_drivers = quantum.openstack.common.notifier.rabbit_notifier
 lock_path = /var/lock/quantum
 # Ensure that netns cleanup operations kill processes and remove ports
-force = true
+# force = true
 [AGENT]
 root_helper = sudo /usr/bin/quantum-rootwrap /etc/quantum/rootwrap.conf
 [QUOTAS]
\ No newline at end of file