The WebLogic Kubernetes Operator enables you to override some of the domain configuration using configuration overrides (also called situational configuration). See supported overrides. Overriding the tuning parameters such as MaxMessageSize and PAYLOAD, for Oracle SOA Suite domains is not supported using the configuration overrides feature. However, you can override them using the following steps:
Specify the new value using the environment variable K8S_REFCONF_OVERRIDES
in serverPod.env
section in domain.yaml
configuration file (example path: <domain-creation-output-directory>/weblogic-domains/soainfra/domain.yaml
) based on the servers to which the changes are to be applied.
For example, to override the value at the Administration Server pod level:
spec:
adminServer:
serverPod:
env:
- name: K8S_REFCONF_OVERRIDES
value: "-Dweblogic.MaxMessageSize=78787878"
- name: USER_MEM_ARGS
value: '-Djava.security.egd=file:/dev/./urandom -Xms512m -Xmx1024m '
serverStartState: RUNNING
For example, to override the value at a specific cluster level (soa_cluster
or osb_cluster
):
- clusterName: soa_cluster
serverService:
precreateService: true
serverStartState: "RUNNING"
serverPod:
env:
- name: K8S_REFCONF_OVERRIDES
value: "-Dsoa.payload.threshold.kb=102410"
Note: When multiple system properties are specified for
serverPod.env.value
, make sure each system property is separated by a space.
Apply the updated domain.yaml
file:
$ kubectl apply -f domain.yaml
Note: The server pod(s) will be automatically restarted (rolling restart).
In an Oracle SOA Suite environment deployed using the operator, accessing Deployments from the WebLogic Server Administration Console home page may display the error message Unexpected error encountered while obtaining monitoring information for applications
. This error does not have any functional impact and can be ignored. You can verify that the applications are in Active state from the Control tab in Summary of deployments page.
In an Oracle SOA Suite environment deployed using the operator, the Enterprise Manager Console may intermittently display the following error when the domain servers are restarted:
ADF_FACES-30200: For more information, please see the server's error log for an entry beginning with: The UIViewRoot is null. Fatal exception during PhaseId: RESTORE_VIEW 1.
You can refresh the Enterprise Manager Console URL to successfully log in to the Console.
For Oracle SOA Suite composite applications to access the external URLs over the internet (if your cluster is behind a http proxy server), you must configure the following proxy parameters for Administration Server and Managed Server pods.
-Dhttp.proxyHost=www-your-proxy.com
-Dhttp.proxyPort=proxy-port
-Dhttps.proxyHost=www-your-proxy.com
-Dhttps.proxyPort=proxy-port
-Dhttp.nonProxyHosts="localhost|soainfra-adminserver|soainfra-soa-server1|soainfra-osb-server1|...soainfra-soa-serverN|*.svc.cluster.local|*.your.domain.com|/var/run/docker.sock"
To do this, edit the domain.yaml
configuration file and append the proxy parameters to the spec.serverPod.env.JAVA_OPTIONS
environment variable value.
For example:
serverPod:
env:
- name: JAVA_OPTIONS
value: -Dweblogic.StdoutDebugEnabled=false -Dweblogic.ssl.Enabled=true -Dweblogic.security.SSL.ignoreHostnameVerification=true -Dhttp.proxyHost=www-your-proxy.com -Dhttp.proxyPort=proxy-port -Dhttps.proxyHost=www-your-proxy.com -Dhttps.proxyPort=proxy-port -Dhttp.nonProxyHosts="localhost|soainfra-adminserver|soainfra-soa-server1|soainfra-osb-server1|...soainfra-soa-serverN|*.svc.cluster.local|*.your.domain.com|/var/run/docker.sock"
- name: USER_MEM_ARGS
value: '-Djava.security.egd=file:/dev/./urandom -Xms256m -Xmx1024m '
volumeMounts:
Note: The
-Dhttp.nonProxyHosts
parameter must have the pod names of the Administration Server and each Managed Server. For example:soainfra-adminserver
,soainfra-soa-server1
,soainfra-osb-server1
, and so on.
Apply the updated domain.yaml
file:
$ kubectl apply -f domain.yaml
Note: The server pod(s) will be automatically restarted (rolling restart).
In an Oracle SOA Suite domain deployed including the Oracle Enterprise Scheduler (ESS) component, the following ESS WebServices WSDL URLs shown in the table format in the ess/essWebServicesWsdl.jsp
page are not reachable outside the Kubernetes cluster.
ESSWebService
EssAsyncCallbackService
EssWsJobAsyncCallbackService
Follow these steps to configure the external access for the Oracle Enterprise Scheduler WebServices WSDL URLs:
http://<LOADBALANCER-HOST>:<port>/console
domain1.example.com
.30305
.30443
.Note: Do not restart servers from the Administration Console.
In an Oracle SOA Suite domain environment upgraded to the release 21.1.2, some gif images are not rendered in the Oracle Serice Bus console pipeline configuration page, as their corresponding url paths are not exposed via the Ingress path rules in the earlier releases (for Non-SSL and SSL termination). To resolve this issue, perform the following steps to apply the latest ingress configuration:
$ cd ${WORKDIR}
$ helm upgrade <helm_release_for_ingress> \
charts/ingress-per-domain \
--namespace <domain_namespace> \
--reuse-values
Note:
helm_release_for_ingress
is the ingress name used in the corresponding helm install command for the ingress installation.
For example, to upgrade the NGINX
based ingress configuration:
$ cd ${WORKDIR}
$ helm upgrade soa-nginx-ingress \
charts/ingress-per-domain \
--namespace soans \
--reuse-values
See the general frequently asked questions for using the WebLogic Kubernetes Operator.