-
Type:
Bug
-
Status: Resolved
-
Priority:
Minor
-
Resolution: Fixed
-
Affects Version/s: None
-
Fix Version/s: explorer-22.0.0
-
Component/s: Explorer
-
Team:PLATFORM
-
Sprint:nxplatform #33
-
Story Points:1
Sometimes, the reference explorer export build fails because the nuxeo pod cannot get scheduled on any node:
Warning FailedScheduling 54s (x5 over 4m24s) default-scheduler 0/34 nodes are available: 1 node(s) were unschedulable, 14 node(s) had taints that the pod didn't tolerate, 19 Insufficient cpu. Normal NotTriggerScaleUp 35s (x27 over 5m44s) cluster-autoscaler pod didn't trigger scale-up (it wouldn't fit if a new node is added): 5 node(s) had taints that the pod didn't tolerate, 1 max node group size reached
The export Helm chart is relying on the ~1.2.1 nuxeo chart, which doesn't take into account Kubernetes tolerations and nodeSelector, thus the related pods cannot be explicitly scheduled on the platform dedicated node pool, which has a sufficient amount of nodes and CPU.
Let's use the latest nuxeo Helm chart: 3.0.3.
- is related to
-
NXP-30377 Add tolerations/nodeSelector to nuxeo Helm chart 1.0.x
-
- Resolved
-
- Is referenced in
