-
Type:
Sub-task
-
Status: In Progress
-
Priority:
High
-
Resolution: Unresolved
-
Fix Version/s: 4.0-rc
-
Component/s: Legacy/Streaming and Messaging, Test/benchmark
-
Labels:None
-
Change Category:Operability
-
Complexity:Normal
-
Platform:All
-
Impacts:None
Let's load a cluster up with data and start killing nodes. We can do hard failures (node terminations) and soft failures (process kills) We plan to observe the following:
- Can nodes successfully bootstrap?
- How long does it take to bootstrap
- What are the effects of TLS on and off (e.g. on stream time)
- Are hints properly played after a node restart
- Do nodes properly shutdown and start back up.