Hard/soft Timeout
Issue using soft Timeout with replica
Current implementation with replica 1:
Current implementation with replica 2. Deploy and undeploy messages are send using normal topic, only one replica will be fetch them.
Solution with Hard Timeout with replica
Hard Timeout by AC-Runtime. the instance in Timeout could be send by ACM-Runtime as sync message; all participant receive sync messages.
Make a configurable timeout for operations in ACM-R to Participants
Currently ACM-Runtime fetches "maxStatusWaitMs" from properties file, uses that value to calculate the timeout for all operations.
What can be implemented:
- Add more specific value for each operation in properties file as default
- Add specific value for each operation in composition definition that override the default value
application.yaml
runtime: topics: operationTopic: policy-acruntime-participant syncTopic: acm-ppnt-sync participantParameters: heartBeatMs: 20000 maxStatusWaitMs: 200000 deployTimeoutMs: 1000000 undeployTimeoutMs: 200000 updateTimeoutMs: 1000000 migrateTimeoutMs: 1000000 deleteTimeoutMs: 200000
service template
tosca_definitions_version: tosca_simple_yaml_1_3 metadata: deployTimeoutMs: 1000000 undeployTimeoutMs: 200000 updateTimeoutMs: 1000000 migrateTimeoutMs: 1000000 deleteTimeoutMs: 200000