-
Bug
-
Resolution: Unresolved
-
Normal
-
None
-
4.17
-
None
-
False
-
Description of problem:
ClusterIP type service got duplicate IP on running a perfscale workoad.
Version-Release number of selected component (if applicable):
4.17.3 GA
Affected Platforms:
ROSA-HCP
How reproducible:
Frequently
Steps to Reproduce:
- Cluster Density-v2 workload on 90 node ROSA HCP
- At higher QPS and Burst, 50/50
- With limited KAS Max-request-inflight 300 & max-mutating-request-inflight 100
- Pod stuck at readiness check on service
Actual results:
Duplicate IP
# oc describe svc -n cluster-density-v2-512 cluster-density-1
Name: cluster-density-1
Namespace: cluster-density-v2-512
Labels: kube-burner-index=2
kube-burner-job=cluster-density-v2
kube-burner-runid=a6060414-8f64-42b9-a633-8ae38c041c13
kube-burner-uuid=7bc0ae09-b9f7-4fb2-bdf3-252ad1c6a985
Annotations: <none>
Selector: app=nginx
Type: ClusterIP
IP Family Policy: SingleStack
IP Families: IPv4
IP: 172.30.152.126
IPs: 172.30.152.126
Port: http 80/TCP
TargetPort: 8080/TCP
Endpoints: 10.128.46.67:8080,10.128.50.66:8080,10.128.72.68:8080 + 3 more...
Session Affinity: None
Events:
Type Reason Age From Message
---- ------ ---- ---- -------
Warning ClusterIPAlreadyAllocated 9m57s (x90 over 4h38m) ipallocator-repair-controller Cluster IP [IPv4]:172.30.152.126 was assigned to multiple services; please recreate service
Warning ClusterIPAlreadyAllocated 6m58s (x90 over 4h34m) ipallocator-repair-controller Cluster IP [IPv4]:172.30.152.126 was assigned to multiple services; please recreate service
Already assigned to a different service
# oc get svc -A | grep 172.30.152.126 cluster-density-v2-400 cluster-density-3 ClusterIP 172.30.152.126 <none> 80/TCP 4h49m cluster-density-v2-512 cluster-density-1 ClusterIP 172.30.152.126 <none> 80/TCP 4h48m
Expected results:
Unique service IP
