problem phenomenon：
NAME READY STATUS RESTARTS AGE
kuryr-cni-ds-spq5q 0/1 CrashLoopBackOff 233 16h
kuryr-controller-666658496d-hbl4t 0/1 Running 106 16h
tomcat-5cf6fccd5d-5jmh6 0/1 ContainerCreating 0 15h
# kubectl get kp tomcat-5cf6fccd5d-5jmh6 -oyaml |egrep 'interface_id:|active'
active: false
interface_id: d223e87d-6c38-45d3-bbe3-af35ea439a60
# openstack port show d223e87d-6c38-45d3-bbe3-af35ea439a60 |grep ' status'
| status | DOWN |
The ovs port up event is lost, in fact the state of the ovs-port is always up and functional, but the state of the neutron port is down, in this case an exception that exceeds the number of retries will cause cni to reboot all the time, and this exception and reboot does not really solve the corresponding port's pod.
It is possible that the state of a particular neutron port port is down, but it is not what we would expect if this causes the entire kuryr component to keep rebooting exceptionally.
Is it possible that cni will continue to add processes after more than the number of retries, even if the port is not up?
Note：
Batch creation of pods on a kuryr-node, port up events are missing for a few ports (the ovs-agent in my environment had an exception in the middle and the port up events were not notified to the neutron-server).
kuryr-cni daemon crash log:
ERROR kuryr_kubernetes.cni.daemon.service [-] Error when processing addNetwork request. CNI Params: {'CNI_IFNAME': 'eth0', 'CNI_NETNS': '/proc/3256304/ns/net', 'CNI_PATH': '/opt/cni/bin', 'CNI_COMMAND': 'ADD', 'CNI_CONTAINERID': '0fe17cdbc5ff678b01d797cf8a4e7a4a027356b987681aadc563260fde1a9840', 'CNI_ARGS': 'IgnoreUnknown=1;K8S_POD_NAMESPACE=ljx;K8S_POD_NAME=tomcat-5cf6fccd5d-5jmh6;K8S_POD_INFRA_CONTAINER_ID=0fe17cdbc5ff678b01d797cf8a4e7a4a027356b987681aadc563260fde1a9840'}: BrokenPipeError: [Errno 32] Broken pipe
ERROR kuryr_kubernetes.cni.daemon.service Traceback (most recent call last):
ERROR kuryr_kubernetes.cni.daemon.service File "/usr/local/lib/python3.6/site-packages/kuryr_kubernetes/cni/daemon/service.py", line 82, in add
ERROR kuryr_kubernetes.cni.daemon.service vif = self.plugin.add(params)
ERROR kuryr_kubernetes.cni.daemon.service File "/usr/local/lib/python3.6/site-packages/kuryr_kubernetes/cni/plugins/k8s_cni_registry.py", line 83, in add
ERROR kuryr_kubernetes.cni.daemon.service d = self.registry[kp_name]
ERROR kuryr_kubernetes.cni.daemon.service File "<string>", line 2, in __getitem__
ERROR kuryr_kubernetes.cni.daemon.service File "/usr/lib64/python3.6/multiprocessing/managers.py", line 756, in _callmethod
ERROR kuryr_kubernetes.cni.daemon.service conn.send((self._id, methodname, args, kwds))
ERROR kuryr_kubernetes.cni.daemon.service File "/usr/lib64/python3.6/multiprocessing/connection.py", line 210, in send
ERROR kuryr_kubernetes.cni.daemon.service self._send_bytes(_ForkingPickler.dumps(obj))
ERROR kuryr_kubernetes.cni.daemon.service File "/usr/lib64/python3.6/multiprocessing/connection.py", line 408, in _send_bytes
ERROR kuryr_kubernetes.cni.daemon.service self._send(header + buf)
ERROR kuryr_kubernetes.cni.daemon.service File "/usr/lib64/python3.6/multiprocessing/connection.py", line 372, in _send
ERROR kuryr_kubernetes.cni.daemon.service n = write(self._handle, buf)
ERROR kuryr_kubernetes.cni.daemon.service BrokenPipeError: [Errno 32] Broken pipe
ERROR kuryr_kubernetes.cni.daemon.service
INFO werkzeug [-] 127.0.0.1 - - [21/Oct/2020 02:31:48] "POST /addNetwork HTTP/1.1" 500 -
DEBUG cotyledon._service_manager [-] Shutdown finish _shutdown /usr/local/lib/python3.6/site-packages/cotyledon/_service_manager.py:320
kuryr-controller dead log：
ERROR kuryr_kubernetes.handlers.logging Traceback (most recent call last):
ERROR kuryr_kubernetes.handlers.logging File "/usr/local/lib/python3.6/site-packages/kuryr_kubernetes/handlers/logging.py", line 37, in __call__
ERROR kuryr_kubernetes.handlers.logging self._handler(event, *args, **kwargs)
ERROR kuryr_kubernetes.handlers.logging File "/usr/local/lib/python3.6/site-packages/kuryr_kubernetes/handlers/retry.py", line 93, in __call__
ERROR kuryr_kubernetes.handlers.logging self._handler.set_liveness(alive=False)
ERROR kuryr_kubernetes.handlers.logging File "/usr/local/lib/python3.6/site-packages/oslo_utils/excutils.py", line 220, in __exit__
ERROR kuryr_kubernetes.handlers.logging self.force_reraise()
ERROR kuryr_kubernetes.handlers.logging File "/usr/local/lib/python3.6/site-packages/oslo_utils/excutils.py", line 196, in force_reraise
ERROR kuryr_kubernetes.handlers.logging six.reraise(self.type_, self.value, self.tb)
ERROR kuryr_kubernetes.handlers.logging File "/usr/local/lib/python3.6/site-packages/six.py", line 703, in reraise
ERROR kuryr_kubernetes.handlers.logging raise value
ERROR kuryr_kubernetes.handlers.logging File "/usr/local/lib/python3.6/site-packages/kuryr_kubernetes/handlers/retry.py", line 78, in __call__
ERROR kuryr_kubernetes.handlers.logging self._handler(event, *args, **kwargs)
ERROR kuryr_kubernetes.handlers.logging File "/usr/local/lib/python3.6/site-packages/kuryr_kubernetes/handlers/k8s_base.py", line 84, in __call__
ERROR kuryr_kubernetes.handlers.logging self.on_present(obj)
ERROR kuryr_kubernetes.handlers.logging File "/usr/local/lib/python3.6/site-packages/kuryr_kubernetes/controller/handlers/kuryrport.py", line 88, in on_present
ERROR kuryr_kubernetes.handlers.logging self._drv_vif_pool.activate_vif(data['vif'])
ERROR kuryr_kubernetes.handlers.logging File "/usr/local/lib/python3.6/site-packages/kuryr_kubernetes/controller/drivers/vif_pool.py", line 1258, in activate_vif
ERROR kuryr_kubernetes.handlers.logging self._vif_drvs[vif_drv_alias].activate_vif(vif)
ERROR kuryr_kubernetes.handlers.logging File "/usr/local/lib/python3.6/site-packages/kuryr_kubernetes/controller/drivers/vif_pool.py", line 169, in activate_vif
ERROR kuryr_kubernetes.handlers.logging self._drv_vif.activate_vif(vif)
ERROR kuryr_kubernetes.handlers.logging File "/usr/local/lib/python3.6/site-packages/kuryr_kubernetes/controller/drivers/neutron_vif.py", line 92, in activate_vif
ERROR kuryr_kubernetes.handlers.logging raise k_exc.ResourceNotReady(vif)
ERROR kuryr_kubernetes.handlers.logging kuryr_kubernetes.exceptions.ResourceNotReady: Resource not ready: VIFOpenVSwitch(active=False,address=fa:16:3e:b6:d2:4a,bridge_name='br-int',has_traffic_filtering=True,id=8d85b962-c75a-4348-b68c-6ff5c0a951a6,network=Network(84f65523-4252-4197-85ec-75e10dc46389),plugin='ovs',port_profile=VIFPortProfileOpenVSwitch,preserve_on_delete=False,vif_name='tap8d85b962-c7')
ERROR kuryr_kubernetes.handlers.logging
ERROR kuryr_kubernetes.controller.managers.health [-] Component KuryrPortHandler is dead.
INFO kuryr_kubernetes.controller.service [-] Service 'KuryrK8sService' stopping
INFO kuryr_kubernetes.watcher [-] Stopped watching '/apis/openstack.org/v1/kuryrports'
INFO kuryr_kubernetes.watcher [-] Stopped watching '/apis/openstack.org/v1/kuryrnetworks'
INFO kuryr_kubernetes.watcher [-] Stopped watching '/api/v1/namespaces'
INFO kuryr_kubernetes.watcher [-] Stopped watching '/api/v1/pods'
INFO kuryr_kubernetes.watcher [-] Stopped watching '/apis/networking.k8s.io/v1/networkpolicies'
INFO kuryr_kubernetes.watcher [-] No remaining active watchers, Exiting...
INFO kuryr_kubernetes.controller.service [-] Service 'KuryrK8sService' stopping
INFO kuryr_kubernetes.controller.service [-] Service 'KuryrK8sService' stopped
Similar bug：
https:/