Presentation is loading. Please wait.

Presentation is loading. Please wait.

Tomi Juvonen Software Architect, Nokia

Similar presentations


Presentation on theme: "Tomi Juvonen Software Architect, Nokia"— Presentation transcript:

1 Tomi Juvonen Software Architect, Nokia
Telco Scheduled Host Maintenance POC Tomi Juvonen Software Architect, Nokia

2 OPNFV Doctor Maintenance Use Case
4. Switch to SBY configuration V Administrator Consumer C1 Consumer C2 Consumer C3 3. Maintenance Notification (VM ID) 1. Maintenance Request (Server S3) 5. Instruction (VM ID) OpenStack Northbound Interface Virtualized Infrastructure Manager (VIM), e.g. OpenStack VM-1 VM-2 VM-7 VM-4 6. Execute Instruction - e.g. migrate VM Resource Map Server – VM mapping Server S1 VM-1, VM-2 Server S2 VM-7 Server S3 VM-4 Resource Pool Hypervisor Hypervisor Hypervisor Ownership information VM-1, VM-7 Consumer C1 VM-2 Consumer C2 VM-4 Consumer C3 Hardware Server S1 Hardware Server S2 Hardware Server S3 2. Which VMs are affected? Find Consumer owning the VM(s) from the database.

3 Maintenance Design http://docs. opnfv
Admin Application Migrate, Delete,… Ack + Action Optionally Re-instantiate (and delete) App Manager Schedule maintenance Create Alarm Switch over Cloud configuration DB HAPP1 act NONHAPP1 HAPP2 stdby NONHAPP2 HAPP1 stdby NONHAPP1 Project maintenance alarm Fleet management tool Action Maintenance state notifications Craton Virtualized Infrastructure (Resource Pool) Physical host Physical host Physical host Notifier Workflow engine ??? Alarm Conf. Create Alarm Ceilometer /Aodh Maintenance Empty Admin maintenance alarm Inspector Failure Policy Controller Resource Map Migrate Controller Controller Vitrage Optionally Delete Nova Congress Audit and maintenance workflow actions Migrate, Delete,… Admin action Project action OpenStack project Cloud Infra Entity

4 Demo setup Maintenance demo setup where a single compute node at a time will be emptied for cloud admin maintenance. There is one empty host where application payload is moved in interaction with application manager. Application may do upgrade at the same time. 4 Compute nodes with 48 VCPUs each: overcloud-novacompute-0.opnfvlf.org overcloud-novacompute-1.opnfvlf.org overcloud-novacompute-2.opnfvlf.org overcloud-novacompute-3.opnfvlf.org 2 APPs: ha_app: 2 instances taking 32 VCPUs each. Floating IP is set for the active instance. Uses AntiAffinity so they are assured to be in different host: ha_app1, ha_app2 nonha_app: 5 instances taking 16 VCPUs each: nonha_app1, nonha_app2, nonha_app3, nonha_app4, nonha_app5

5 Maintenance Demo Design
Admin Maintenance Demo Design Schedule maintenance for a host aggregate/cell app_manager inspector maint. workflow Craton Application Event to project about their payload affected and allowed admin actions (migrate, delete or none) Audit Nova for hosts, projects and their payload. Disable compute nodes. Switch over Re-instantiate Project own actions ha_app1 act nonha_app1 ha_app2 stdby nonha_app2 ha_app1 stdby nonha_app1 Answer admin actions Migrate: ha_app1 Delete: nonha_app1 Virtualized Infrastructure (Resource Pool) Physical host Physical host Physical host Migrate: ha_app1 Delete: nonha_app1 Admin actions Host empty, Ready for maintenance Empty Disable host automatic fault management Maintenance Migrate Actual host maintenance done here (out of demo scope). After that compute back to enabled Delete Enable host automatic fault management Maintenance done Admin action Project action

6 Messaging in POC Maintenance notification/event alarm to project:
Event type: maintenance.scheduled Payload: {'service': 'craton', 'instances': ['nonha_app3', 'ha_app1'], 'allowed_actions': ['migrate', 'delete', 'none'], 'session_id': ' ', 'reply_url': ' 'maintenance_at': ' T10:06:00', 'state': 'scheduled', 'host_id': '77', 'project_id': 'fa0c787d1c2a431f8ce3d58073d08d78', 'metadata': {'openstack_version': 'Pike'}} Project needs to know which of his server Instances are affected by maintenance and what are the possible admin actions. Project replies (In Demo via API) with action per instance: {'ha_app1': 'migrate', 'nonha_app3': 'delete'} Maintenance notification/event alarm to admin: Event type: maintenance.host Payload: {'service': 'craton', 'host': 'overcloud-novacompute-0.opnfvlf.org', 'project_id': 'fa0c787d1c2a431f8ce3d58073d08d78', 'state': 'in_maintenance', 'session_id': ' '} Admin level needs to know about ongoing maintenance and when it is over. For example Inspector needs to disable fault management during maintenance.

7 Thanks!!!


Download ppt "Tomi Juvonen Software Architect, Nokia"

Similar presentations


Ads by Google