Details
-
Enhancement
-
Resolution: Unresolved
-
Major
-
None
-
TEMPLATE 1.4.10.GA
-
Interactive Demo/Tutorial, User Experience
-
CLOUD Maintenance Sprint 28
Description
When using a persistent storage, RH-SSO pod often needs 2-3 restarts till the DB becomes reachable to utilize the services of the MySQL / PostgreSQL DB properly. See details.
The proposal is to use OpenShift init containers to achieve proper pod starting sequence control (ensure RH-SSO pod is started only in moment related MySQL / PostgreSQL pod is able to receive JDBC connections already):
The proposed fix (recognized as working properly for MySQL DB case, needs to be tested for PostgreSQL's DB case yet) is as follows:
@@ -488,6 +488,23 @@
},
"spec": {
"terminationGracePeriodSeconds": 75,
+ "initContainers": [
+ {
+ "name": "wait-for-${APPLICATION_NAME}-mysql",
+ "image": "busybox",
+ "command": [
+ "sh",
+ "-c",
+ "until nc -w 3 -z ${SVC} 3306; do echo waiting for ${SVC} to become available ; sleep 7 ; done"
+ ],
+ "env": [
+ {
+ "name": "SVC",
+ "value": "${APPLICATION_NAME}-mysql"
+ }
+ ]
+ }
+ ],
"containers": [
{
"name": "${APPLICATION_NAME}",
