[Edge case] MCM rollout stuck if rollout + scale-up triggered together
Opened this issue · 1 comments
himanshu-kun commented
How to categorize this issue?
/area robustness
/kind bug
/priority 2
What happened:
There is a case where if
- there are > 1 active machineSets (machineSet with
.spec.replicas
> 0), and - machineDeployment is updated such that it starts referring to a new machineClass , and
machineDeployment.spec.replicas
isincreased
(in decrease case issue doesn't happen)
then mcm starts to panic.
Furthermore , if the panic doesn't happen , the rollout will be stuck because scale-up logic is run before rollout logic (where new machineSet creation happens), and so new machineSet is never created.
Need to solve this in two steps:
- Address the panic situation
- #814
- Allow the rollout to proceed
What you expected to happen:
How to reproduce it (as minimally and precisely as possible):
Anything else we need to know?:
A result of the changes introduced in #765
Environment:
mcm v0.49.0
- Kubernetes version (use
kubectl version
): - Cloud provider or hardware configuration:
- Others:
himanshu-kun commented
t=0 ms1
t=1 ms1 , ms2
t=2 ms1 , ms2 , ms3 (+scale-up)