nfv_client returns None instead of Internal Errors
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
StarlingX |
Fix Released
|
Medium
|
Al Bailey |
Bug Description
Brief Description
-----------------
During a kubernetes rootca update orchestration, the services are restarted.
This causes the vim api to be restarted, and can return a socket error if something is querying
the vim api at that time.
This has the unfortunate side-effect of the API returning successfully, but returns a None value rather than reporting the error.
The bug is in the 'error' reason extraction code. it 'returns' when it is supposed to 'break' from the loop.
(I am referring to a previous branch, since this will soon be fixed in the current branch)
https:/
Severity
--------
Minor
Steps to Reproduce
------------------
Its difficult to reproduce.
The easier way is to loop and repeatedly query the API, while repeatedly restarting the API.
Expected Behavior
------------------
The error should be reported as part of the query.
Actual Behavior
----------------
The query returns None. When performing a GET, this can make the code think the object has been deleted, when in fact, there is an internal error.
Reproducibility
---------------
Rare
System Configuration
-------
Any
Branch/Pull Time/Commit
-------
The bug has been there for over 3 years. Its rare and 'sneaky'
Last Pass
---------
Not Applicable, due to the sneakiness
Timestamp/Logs
--------------
N/A
Test Activity
-------------
Feature Testing
Workaround
----------
Its so random and in-frequent, running the test again will typically work.
Changed in starlingx: | |
assignee: | nobody → Al Bailey (albailey1974) |
Changed in starlingx: | |
importance: | Undecided → Medium |
tags: | added: stx.6.0 stx.nfv |
Fix proposed to branch: master /review. opendev. org/c/starlingx /nfv/+/ 817115
Review: https:/