-
Bug
-
Resolution: Unresolved
-
Normal
-
None
-
4.16, 4.17, 4.17.z, 4.16.z, 4.18, 4.18.z, 4.19.z, 4.19
Description of problem:
extensive log output from default setting bash into set -x mode. That causes a massive amount of logs to be produced without any real benefit stressing collectors and log backends.
Version-Release number of selected component (if applicable):
4.16+
How reproducible:
Always
Steps to Reproduce:
1. Install OpenShift 4.16+
2. check the ovnkube-node pod logs
$ oc -n openshift-ovn-kubernetes get po -l app=ovnkube-node NAME READY STATUS RESTARTS AGE ovnkube-node-464zd 8/8 Running 8 12d ovnkube-node-9rrtb 8/8 Running 8 12d ... $ oc -n openshift-ovn-kubernetes logs $(oc -n openshift-ovn-kubernetes get po -l app=ovnkube-node -oname | head -1) ovn-acl-logging | wc -l 431762
Actual results:
++ tr -s '\t' ' ' ++ cut '-d ' -f1 + file_size=0 + '[' 0 -gt 50000000 ']' ++ ls -1 /var/log/ovn/acl-audit-log.log ++ wc -l + num_files=1 + '[' 1 -gt 5 ']' + sleep 30 + true + '[' -f /var/log/ovn/acl-audit-log.log ']' ++ du -b /var/log/ovn/acl-audit-log.log ...
Expected results:
way less logs
PR upstream submitted as well https://github.com/openshift/cluster-network-operator/pull/2761
