Details
-
Bug
-
Resolution: Done
-
Major
-
7.8.0.GA, 7.9.0.GA, 7.8.1.GA
-
2020 Week 46-48 (from Nov 9)
-
Undefined
-
https://github.com/jboss-container-images/rhpam-7-openshift-image/pull/484, https://github.com/jboss-container-images/rhpam-7-openshift-image/pull/485, https://github.com/jboss-container-images/rhdm-7-openshift-image/pull/378, https://github.com/jboss-container-images/rhdm-7-openshift-image/pull/379, https://github.com/kiegroup/kie-cloud-operator/pull/554, https://github.com/kiegroup/kie-cloud-operator/pull/557
Description
The probes' spec defined for the Business Central (console) is too short and is causing premature POD termination in some cases.
17:26:52,740 WARN [org.appformer.maven.integration.MavenRepository] (Thread-166) Unable to resolve artifact: com.redhat.demos:rules-model:pom:1.0.0-SNAPSHOT *** JBossAS process (1010) received TERM signal ***
Depending on the size of BC's index and internal git the startup may take longer than 30s as defined here https://github.com/kiegroup/kie-cloud-operator/blob/master/config/7.8.1/common.yaml#L59
As the amount of time required to startup BC is not deterministic and depends on the size of internal Git repo and projects, the DC probes should use a long timeout to ensure it can finish the indexing process and avoid a premature SIG TERM which can corrupt the volume file system.