Description of problem: In 3.3.0 images the fluentd is deployed by daemonset instead of dc, so the way to pass throttling settings into fluentd pod is broken since not able to mount the created secret to Fluentd container: https://github.com/openshift/origin-aggregated-logging#throttling-logs-in-fluentd Version-Release number of selected component (if applicable): openshift3/logging-curator 3.3.0 0f4e933a812a openshift3/logging-elasticsearch 3.3.0 b7051c8b66d3 openshift3/logging-kibana 3.3.0 32d276bb46ae openshift3/logging-fluentd 3.3.0 0de35fb6abaa openshift3/logging-auth-proxy 3.3.0 aaa3c8315877 openshift3/logging-deployer 3.3.0 000cdaaa18ea How reproducible: Always Steps to Reproduce: 1. Refer to part "Description of problem" 2. 3. Actual results: Expected results: Additional info:
You are right. Not only should the docs be amended to point to the daemonset instead of the deploymentconfig, but because daemonsets can't be modified, the directions have to indicate deleting the DS and re-creating it with a patch. I'll work on this.
I see with latest logging images on brew registry, the fluentd configmap exists, however, the verification of throttling log effect was blocked by kibana UI not accessible, which is described here: https://bugzilla.redhat.com/show_bug.cgi?id=1360675 Will keep monitoring it. In the mean time, please feel free to set this ON_QA if the development work is considered to be done.
Not sure if https://bugzilla.redhat.com/show_bug.cgi?id=1360675#c12 helps you but if you can get past Kibana, this should be fixed.
Checked on OSE v3.3.0.13 with latest 3.3.0 images: 1. Able to put throttling settings into configmap/logging-fluentd 2. Throttling settings can be delivered to fluentd container successfully 3. Throttling settings worked (identified from Kibana UI) Set this to verified.