Bug 1616171
| Summary: | logging-fluentd requires its own priority class to be scheduled on infra nodes | ||
|---|---|---|---|
| Product: | OpenShift Container Platform | Reporter: | Junqi Zhao <juzhao> |
| Component: | Logging | Assignee: | ewolinet |
| Status: | CLOSED ERRATA | QA Contact: | Anping Li <anli> |
| Severity: | urgent | Docs Contact: | |
| Priority: | unspecified | ||
| Version: | 3.11.0 | CC: | aos-bugs, jcantril, jupierce, rmeggins |
| Target Milestone: | --- | ||
| Target Release: | 3.11.0 | ||
| Hardware: | Unspecified | ||
| OS: | Unspecified | ||
| Whiteboard: | |||
| Fixed In Version: | Doc Type: | Bug Fix | |
| Doc Text: |
Cause: If logging was not in a namespace that began with 'openshift-' Fluentd was not able to use the "system-cluster-critical" priority class.
Consequence: Fluentd would not be able to start up.
Fix: We create a priority class for Cluster Logging and configure Fluentd to use that instead.
Result: Fluentd is able to start up, even if not installed to an 'openshift-*' namespace.
|
Story Points: | --- |
| Clone Of: | Environment: | ||
| Last Closed: | 2018-10-11 07:24:57 UTC | Type: | Bug |
| Regression: | --- | Mount Type: | --- |
| Documentation: | --- | CRM: | |
| Verified Versions: | Category: | --- | |
| oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |
| Cloudforms Team: | --- | Target Upstream Version: | |
| Embargoed: | |||
|
Description
Junqi Zhao
2018-08-15 07:43:42 UTC
Do we need a special provision here or move logging to the default namespace for new deployments of 'openshift-logging' ? @liggitt seems like your name in this area of code. Can you provide some guidance? system-priority level pods are allowed in kube-system and openshift-... namespaces only Modifying this bug to reflect the need of a priority class for scheduling. The original fact the version was wrong I'm certain was related to the deployment timing out. Per #c5, fluent is unable to be scheduled because this is an upgrade where the logging stack still lives in a namespace that is not 'openshift-'. At the recommendation of D.Carr, logging needs a priority class that is below system-priority level: * reason 1 - the restriction for priority class you are hitting is specific to cluster critical * reason 2 - fluentd should be second tier to other cluster components We should create a priority class for cluster-logging. We need this class regardless but is a stop gap until we can migrate deployments to 'openshift-logging' in the 4.0 uprades. Example from avesh: apiVersion: scheduling.k8s.io/v1beta1 kind: PriorityClass metadata: name: high-priority value: 100000 globalDefault: false description: "This priority class for high priority Applications" ref: https://kubernetes.io/docs/tasks/administer-cluster/guaranteed-scheduling-critical-addon-pods/ Move the verified. The logging applications in logging namespace can be upgrade to v3.11. It is using priorityClassName: cluster-logging.
$ oc get ds -o yaml |grep priorityClassName
priorityClassName: cluster-logging
Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory, and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. https://access.redhat.com/errata/RHBA-2018:2652 |