This is old and probably not applicable anymore. In the meantime we have moved to a new version of SEBA and we are targeting baremetal deployments only. This bug was seen especially on virtual PODs.

After a significant amount of time, around 90 minutes, the seba-seba-services-tosca-loader pod has disappeared.

Comparing this with a deployment on a baremetal pod (manually installed on arm-pod8), the corresponding pod has status 'Completed' on the baremetal pod:

ubuntu@iec01:~$ kubectl get pods --all-namespaces | grep -e '\<seba'
default seba-base-kubernetes-tosca-loader-n8tg4 0/1 Completed 5 29h
default seba-fabric-78b4dc9754-ddllg 1/1 Running 0 29h
default seba-fabric-crossconnect-5ff47fcff6-t5flp 1/1 Running 0 29h
default seba-kubernetes-774c8599d8-b262p 1/1 Running 0 29h
default seba-onos-service-66fb88f5d4-85h75 1/1 Running 0 29h
default seba-rcord-594f64db8f-v8nw2 1/1 Running 0 29h
default seba-seba-services-tosca-loader-hc2px 0/1 Completed 5 29h
default seba-volt-6795df5d74-rv2gn 1/1 Running 0 29h
ubuntu@iec01:~$ kubectl logs seba-seba-services-tosca-loader-hc2px
Starting TOSCA loader using httpie version: 0.9.9
Loading: /opt/tosca/010-fixtures.yaml
Created models: 'bridge', 'MyDeployment', 'siterole#tech', 'neutron_port_name', 'c_tag', 'siterole#pi', 'next_hop', 'Private', 'siterole#admin', 'device', 's_tag', 'Public dedicated IPv4', 'Public shared IPv4'
Loading: /opt/tosca/020-rcord-subscriber-service.yaml
Created models: 'service#rcord'
Loading: /opt/tosca/030-volt-service.yaml
Created models: 'service#volt'
Loading: /opt/tosca/040-fabric-crossconnect-service.yaml
Created models: 'service#fabric-crossconnect'
Loading: /opt/tosca/061-onos-service.yaml
Created models: 'service#onos', 'onos_app#segmentrouting', 'onos_app#aaa', 'aaa-config-attr', 'onos_app#sadis', 'onos_app#olt', 'sadis-config-attr', 'onos_app#kafka', 'onos_app#dhcpl2relay', 'onos_app#openflow-base', 'olt-config-attr', 'kafka-config-attr', 'onos_app#netcfghostprovider', 'onos_app#hostprovider', 'onos_app#openflow'
Loading: /opt/tosca/070-fabric-service.yaml
Created models: 'service#fabric'
Loading: /opt/tosca/300-service-graph.yaml
Created models: 'service#fabric-crossconnect', 'service#volt', 'service_dependency#volt_fabric-crossconnect', 'service#fabric', 'service#onos', 'service_dependency#onos_volt', 'service_dependency#onos-fabric_fabric', 'service_dependency#onos_fabric-crossconnect', 'service#rcord', 'constraints', 'service_dependency#rcord_volt'

The same happens to the att-workflow-tosca-loader pod:

ubuntu@iec01:~$ kubectl logs att-workflow-tosca-loader-jxf4v
Starting TOSCA loader using httpie version: 0.9.9
Loading: /opt/tosca/010-fixtures.yaml

http: error: ConnectionError: ('Connection aborted.', RemoteDisconnected('Remote end closed connection without response',)) while doing POST request to URL: http://xos-tosca:9102/run

ubuntu@iec01:~$ kubectl get pods --all-namespaces | grep tosca-loader
default att-workflow-tosca-loader-jxf4v 1/1 Running 3 2m7s
default seba-base-kubernetes-tosca-loader-dsdwl 0/1 CrashLoopBackOff 7 22m
default seba-seba-services-tosca-loader-6bcvz 0/1 CrashLoopBackOff 7 22m

ubuntu@iec01:~$ kubectl logs seba-seba-services-tosca-loader-6bcvz
Starting TOSCA loader using httpie version: 0.9.9
Loading: /opt/tosca/010-fixtures.yaml

http: error: ConnectionError: ('Connection aborted.', RemoteDisconnected('Remote end closed connection without response',)) while doing POST request to URL: http://xos-tosca:9102/run
ubuntu@iec01:~$ kubectl logs seba-base-kubernetes-tosca-loader-dsdwl
Starting TOSCA loader using httpie version: 0.9.9
Loading: /opt/tosca/010-fixtures.yaml

http: error: ConnectionError: ('Connection aborted.', RemoteDisconnected('Remote end closed connection without response',)) while doing POST request to URL: http://xos-tosca:9102/run