- Look for opportunities to simplify the run-jnlp-client code - Confirm why are we setting -Xms, what effect it has and if it is needed - Allow -Xmx=<amount> to be specified as a specific amount, not just a % of container size - Consider capping % of container size calculation for large containers (i.e. min(% * container size, 256MB) - Consider master switch for disabling all smarts - Consider not using JAVA_OPTS as an environment variable name as it might conflict with spawned Java processes - Investigate how to tune based on container *request*, not container limits, e.g. using downward API - Investigate if setting additional environment variable presets would make sense for subsequent workload (JAVA_OPTS, MAVEN_OPTS, GRADLE_OPTS, JAVA_TOOL_OPTIONS, _JAVA_OPTIONS?)
some bugzilla maintenance per request from Paul Weil wrt bugs with no target and 3.7 closing up ... setting this untargetted bug to post 3.7
https://github.com/openshift/jenkins/pull/442
Trello card: https://trello.com/c/SvSBSkOy/1425-8-jenkins-memory-cleanup-evg
Test with openshift3/jenkins-2-rhel7:v3.9(4e3dc986043c) openshift3/jenkins-slave-base-rhel7:v3.9(3b1e124d2bcd) openshift3/jenkins-slave-maven-rhel7:v3.9(aa5c4ce0e097) could move to verified
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-2018:0489