Defect HC-14223
Adding a retry and wait logic in STA stage to avoid crashing data query server
Software
Customer case | Applies to | Corrected in |
---|---|---|
CS0280531 | V9.1.8.0 | 9.1.9.0 |
Observed behavior
The query server crashes when unlimited retries are permitted with no wait times during this cycle:
NiFi Flow > auth.reindex - STA Zookeeper > Find STA > Remove STA from Zookeeper > "POST Query Config Profile"
Expected behavior
There should be retry logic and timeouts configured so that infinite callback loops can be avoided.
Resolution
While trying to validate the businessUser through a call back to tsweb, there is a failure. The code was updated to add a RetryFlowFile to registry/flows/base/DatabaseSTAZookeeperStage1.json, which will retry a limited number of times, with waits, to avoid flooding the query server with additional requests.