vnf monitoring fails after multiple vnf create/delete operations

Bug #1513972 reported by Santosh
This bug affects 2 people
Affects Status Importance Assigned to Milestone
Anshu Kumar

Bug Description

1. Use following template to create a vnfd.

template_name: http-ping
description: http-ping-cirros-example

  Id: cirros-vnfd-http-moniotring
  vendor: canonical
  version: 1

    id: vdu1
    vm_image: cirros-0.3.4-x86_64-uec
    instance_type: m1.small
    user_data_format: RAW
    user_data: |
       cd /tmp
       rm -f
       cat > << EOL
       eth0_ip=$(ifconfig eth0|grep 'inet addr'|awk -F: '{print $2}'| awk '{print $1}')
       while true; do echo -e "HTTP/1.0 200 OK\r\n\r\nWelcome to $eth0_ip" | sudo nc -l -p 80 ; done&
       sudo chmod +x
       sudo ./ &
       sleep 180
       cat > kill << EOL
       cat <<'EOF' > kill
       pid = `ps -ef | grep httpsrv | grep exe | grep -v grep | awk '{print $1}'`
       echo "killing $pid"
       kill -9 $pid
       sudo chmod +x kill

        network: net_mgmt
        management: true

      availability_zone: nova

          monitoring_delay: 60
          retry: 5
          timeout: 10
          port: 80
          failure: respawn

      param0: key0
      param1: key1

2. Create 8 vnfs with above template. Notice, vnf should respawn approximately every 3 minutes.
   Verify all vnfs respawn after every 3 mins.
3. Run the functionality testcases using "tox -e functional" . In functionality test, there are multiple create/delete vnfs .
4. After the tests are run some of vnfs will go to ERROR state (because of lack or resource - assuming only 10 vm instances is maximum possible). At this point you will notice monitoring doesn't happen and vnfs are not restarted at all

Revision history for this message
Santosh (ksantosh-cs) wrote :
Changed in tacker:
importance: Undecided → Medium
Changed in tacker:
assignee: nobody → Anshu Kumar (anshu-choubey)
Revision history for this message
dharmendra (dharmendra-kushwaha) wrote :

I tested this scenario on current master code with user_data/config mentioned in bug description.
But could not reproduce it.

Could you please mention the tacker version, Or please test on master.

Changed in tacker:
status: New → Incomplete
To post a comment you must log in.
This report contains Public information  Edit
Everyone can see this information.

Other bug subscribers