Known and Resolved Issues
|
|
|
This document describes current known issues as well as issues that were known and are now resolved. For information about new and changed functionality in WebLogic Server 9.0, see What's New in WebLogic Server 9.0.
The following sections describe known and resolved issues in WebLogic Server:
|
The Administration Console does not work properly if you make it the default Web application in a Managed Server environment. Note: The Administration Console was redesigned and no longer exhibits this issue. |
|||
|
Information about cached JDBC statements is not displayed on the JDBC Monitoring pages. |
|||
|
Deploying an application requires multiple steps. With the exception of the start and stop operations, which can be invoked outside of an edit session, only a single deployment operation (install, upgrade, or delete) can be performed on a single deployment in an edit session. Re-activate the edit session after each of these operations on each deployment. |
|||
|
You cannot specify a default Web application for a server from the Administration Console. Note: The Administration Console was redesigned and no longer exhibits this issue. |
|||
|
When you use the FullSecurityDelegation setting, an empty Note: The Administration Console was redesigned and no longer exhibits this issue. |
|||
|
The Administration Console does not create a new Note: The Administration Console was redesigned and no longer exhibits this issue. |
|||
|
Modifications to the Monitoring tables are not available. Note: The Administration Console was redesigned and no longer exhibits this issue. |
|||
|
The Administration Console uses the JMX management interfaces to interact with WebLogic Server. WebLogic Server supports several MBean servers that provide access to management features from different perspectives. Although these MBean Servers may be disabled in the WebLogic Server configuration files, the Administration Console requires these MBean Servers during operation and will not run successfully without them. |
|||
|
The initial start of the Administration Console on an individual server instance is slow. |
|||
|
The DefaultWebAppContextRoot option for Web servers is not yet implemented. |
|||
|
The Administration Console still shows the default table size even after the preferred table size is set to something other than the default size. |
|||
|
You can not change the order of security providers using the Administration Console. |
|||
|
After a pageflow completes in the Administration Console, it forwards to a different page, typically a table. Pressing the browser back button at this point results in an attempt to load the last JSP file in the completed assistant. At this point, all of the context for this assistant is discarded. BEA recommends that you do not use the browser back button to step back into an assistant once changes are cancelled or finished, and that you do not go back to a previous step in an assistant. Instead, use the navigation links and buttons in the Administration Console. |
|||
|
Refreshing the page while using the Start/Stop option in the Web application assistant produces a page flow error. |
|||
|
If you deploy an 8.1 Web Service, the Administration Console's deployment table displays the Web Application that the Web Service is packaged in and the Web Service. If you click the Web Application name, the Administration Console displays Security Roles and Policies tabs. However, there are no Security Roles and Policies tabs for the Web Service itself. Create a new Web Service for 9.0. The Roles and Policies tabs will be displayed for it in the Administration Console. |
|||
|
On the Domain Logging page, if the limit of 100 logs is reached, additional logs are not visible. |
|||
|
A predicate for a security policy cannot be edited through the Administration Console. |
|||
|
You can use the Administration Console to shut down the Administration Server, but as the Administration Console attempts to refresh itself, it often encounters a problem displaying the page because the Administration Server is no longer available to service requests. This failure manifests itself in different ways, depending on the timing, machine load, and other deployments. |
|||
|
A JMS Bridge destination requires a Connection Factory JNDI Name and a Destination JNDI Name. If the Destination JNDI Name field is not specified, all bridge destinations are considered one single physical destination, and the following message is displayed:
|
|||
|
An EJB-style Web service can not be tested through the Administration Console. The EJB Module-->Testing tab cannot be used to test the deployment of an EJB Module (JAR file). |
|||
|
Starting or restarting a Web application does not appear in the Change List in the Change Center. |
|||
|
Currently, no more than 50 users or groups from an external LDAP server or database can be displayed in the Administration Console. |
|||
|
Currently, you can not use wildcards in the Administration Console to restrict the search of users and groups. |
|||
|
After adding security providers to a security realm, the new security providers appear in the Change list and the Change list will indicate which servers need to be rebooted to add the providers to the realm. |
|||
|
When you upgrade a 7.x |
|||
|
The Dump Thread Stacks feature in the Administration Console (Server page --> Monitoring --> Performance --> Lock&Edit --> Dump Thread Stacks) only works if the server is running on JRockit. Thread dumps work on all Virtual Machines if you use the WebLogic Scripting Tool. |
|||
|
The Administration Console can not be used to delete a certificate from the certificate registry. |
|||
|
Using the Copy button when copying a registered certificate to a file causes an error message to be displayed in the Administration Console. |
|||
|
The information displayed in the Change Center after a deployment action (install, start, stop, delete) is confusing. |
|||
|
Applications deployed on a cluster do not appear in the list of deployments for each server belonging to the cluster in the Deployments tab for the Settings of the server. |
|||
|
Message-Driven Beans (MDB) may specify adapter-jndi-name in the deployment descriptor to indicate that the MDB is receiving messages from a resource adapter rather than from JMS. Such a binding entails configuring an ActivationSpec object which is passed to the RA during deployment. The Administration Console does not currently provide a way to view or configure the Resource Adapter that an MDB is bound to. The Administration Console does not currently provide a way to view or configure the ActivationSpec properties on the MDB. |
|||
|
In resource adapters, JNDI names are used for the following three types of objects: The Administration Console should display them all, but does not display the Resource Adapter bean JNDI name. |
|||
|
On the Cluster -> Configuration -> Migration page, you must select the Candidate Machines that you want to use for Migratable Servers at the same time that you specify a cluster-wide data source in the Data Source for Automatic Migration field. |
|||
|
The Deployment Plan and Dependencies pages are incomplete. As a result, you cannot view or edit bindables in a deployment plan through the Administration Console. |
|||
|
JMS service migration does not work using the Administration Console. A JMS server can be successfully migrated using the |
|||
|
The Administration Console sends log messages to the Administration Server log, but there is no way to filter the messages or use the log configuration pages to reconfigure Console logging. |
|||
|
JMS servers cannot be targeted to a migratable target server using the JMS Server -> Target & Deploy page of the Administration Console. Create a persistent file store and target it to a migratable server. Create a JMS server without selecting a persistent store, and target it to a non-migratable server. Reopen the JMS server and modify it to use the persistent file store created in step 1. Target the JMS Server to the same Migratable server that the persistent file store is targeted to. |
|||
|
When a role or policy condition is created or modified in the Administration Console, you must click the. Save button before leaving the role/policy editor page. Otherwise, all the changes made during the creation or modification of the condition(s) that are displayed get lost when you return to the same page. No warning pop-up or message appears to tell you that leaving the page without clicking Save results in losing all changes made. |
|||
|
The ability to configure and monitor durable subscribers was available through the WebLogic Server 8.1 Administration Console, but is not currently available through the WebLogic Server 9.0 Administration Console. |
|||
|
The Administration Console does not provide a way to set control security when custom roles and policies are set. |
|||
|
The Administration Console presents JMS configuration values even if an Message-Driven Bean (MDB) is sourced by a resource adapter. The Administration Console does not provide a way to configure (read and/or write) the adapter-jndi-name for an RA-MDB. |
|||
|
After you enable the administration port from the Administration Console and click the Activate button, the Administration Console is not reachable until the URL used to communicate with the Administration Console is changed to HTTPS and the administration port number. Because the administration port setting is dynamic, the Administration Console should redirect after activating any changes regarding the use of the administration port. |
|||
|
As part of the JMS Modules > [module] > [topic] > Monitoring > Durable Subscribers > Monitoring tab, the table of durable subscribers should, but does not currently, provide a link to browse the messages queued for the subscriber. The New, Move and Delete operations should be, but are not currently, available for selected messages. |
|||
|
If you upload a non-ascii name application in the Administration Console, the file name appears garbled. |
|||
|
After you specify the location for the new plan for a resource adapter, the Administration Console takes you back to the "Settings for <adapter name>" screen where it says that no Deployment Plan is selected. You must have separate directories for the application and the plan instead of placing both files in the same directory. |
|||
|
As administrators configure applications in the Administration Console, some of those configuration changes are saved as deployment plan variables. The Deployment Plan pages allow you to view the variables that have been set in the deployment plan and to delete one or more variables from that plan. The Deployment Plan pages used to view and edit variables are only available for an application, but should also be available for stand-alone deployments. Administrators should navigate to the individual configuration pages for a deployment, module, or component, viewing and editing configuration settings that are specific to that deployment. Deployment plans may also be edited outside of the Administration Console environment by using an XML or text editor. |
|||
|
The Jolt Connection Pool monitoring pages in the Administration Console cannot access the Jolt Runtime MBeans, and therefore Jolt connection pools cannot be monitored or reset using the Administration Console. Use WebLogic Scripting Tool (WLST) access the Jolt Runtime MBean. |
|||
|
The following attributes are currently missing from the Domain > Configuration > JTA page under Advanced:
Sets the maximum allowed duration (in milliseconds) of XA calls to XA resources. This setting applies to the entire domain.
The maximum duration (in milliseconds) that an XA resource is marked as unhealthy. After this duration, the XA resource is declared available again, even if the resource is not explicitly re-registered with the transaction manager. This setting applies to the entire domain.
Maximum number of concurrent requests to resources allowed for each server in the domain. |
|||
|
The Administration Console can be used to attach a policy to a Web Service. However, if you attach more than one policy to a Web Service operation, you receive a |
|||
|
The Diagnostics/ Log Files pages in the Administration Console default to show only the last ten log file entries recorded in the last five minutes. These defaults may be changed by customizing the individual log tables to reflect your preferences. If you do not set preferences and no entries have been recorded in the past five minutes, you may see empty tables because no log entries match this default filtering criteria. Because log files can grow quite large, we limit the size of the log tables in the Administration Console viewer to the most recent 1000 entries that match the filtering criteria. Selecting "ALL" for the maximum results returned on the log table customization pages includes the most recent 1000 entries. |
|||
|
The chooser control is used in several pages in the Administration Console and presents a list of available items, chosen items, and controls to move items from one list to the other. This control does not support moving more than one item at a time using a CTRL+mouse click interaction. |
|||
|
The Web Services Description Language Compiler hangs when non-ascii characters are used in complexType. |
|||
|
If an installed JDBC module has a password-encrypted field encrypted with the wrong key, an exception occurs in the Administration Console. This can occur when you copy a JDBC module from one domain to another. To re-set the password correctly, use |
|||
|
When using the Summary of Asynchronous Tasks function in the Administration Console, invoking Cancel on a running task results in a |
|||
|
If the Administration Console has been modified through a console extension, images and CSS pages get cached by the browser. You may notice the caching when using the MedRec Console. |
|||
|
The Administration Console demands an Administration port for the Suspend and Force Suspend buttons, even though the suspend and force suspend operations do not strictly require it. |
|||
|
When you create a security provider (any provider such as PKI credential mapper, auditing provider, authentication provider, etc.), the creation assistant takes the name of the provider and then returns to that provider's configuration summary page showing the newly added entry. However, the creation is not yet complete. The provider-specific details must be entered for it to be a valid provider configuration. Assistants should result in a complete configuration of a provider, but they currently do not. |
|||
|
The JMX MBean server only allows security management operations on the runtime bean tree. The Administration Console also prevents security management operations while a pending configuration change is in progress. For example, if an authentication provider configuration is changed, the Administration Console prevents creating a new user for that authentication provider until the configuration change is complete, which typically requires a server re-boot. |
|||
|
Mail encoding is currently not configurable on the SMTP configuration page in the diagnostic module. Upon SMTP notification, the mail body and subject are encoded in JVM default encoding. On Japanese Windows and UNIX, JVM default encoding is not mail-safe, which causes various problems such as character garbling, delivery failure, etc. |
|||
|
Using the Administration Console search pain results in a 404 error when it is run on other than the original locale. For example, the 404 error occurs if a French user uses the English version on the French locale. |
|
Netui databinding expression language fails when BeanInfo overrides JavaBean property names. Evaluation of expressions in the Netui databinding expression language throws errors when BeanInfo netui databinding EL does not to work when java.beans.BeanInfo is used to override JavaBean property names. Do not override JavaBean property names using BeanInfo if you use a JavaBean in Netui databinding. If you must databind to such a bean, wrap the bean in a UI version of the bean that follows the JavaBean property naming convention. |
|||
|
The service control generator does not support implicitly wrapped WSDL types. In the 8.1 release of WebLogic Server, service controls generated from wrapped style WSDLs hid the wrapping details from the user, resulting in implicitly wrapped types. In the current 9.0 release implicitly wrapped types are not supported, that is, the generated control does not hide that parameters are wrapped when the service is called. |
|||
|
The Beehive Web Service Control is not included in the version of Beehive shipped with Weblogic Workshop. Use the Web Service Control included as part of WebLogic Server instead. For details and instructions, see the online documentation at http://e-docs.bea.com/workshop/docs81/doc/en/workshop/guide/controls/service/navServiceControl.html. |
|||
|
The version of Beehive shipped with WebLogic Server 9.0 does not include the Web Service metadata sub-project. Consequently, the Web Service metadata documentation has not been included and hyperlinks to the removed documentation content are broken. |
|||
|
Weblogic 9.0 ships with Beehive v1m1. This is an upgrade of the Beehive v1m1 available from the Apache Web site. The following improvements were made: Fixed Apache Beehive issue 829, "Controls deserialization broken because it's looking for the wrong delimiter." See http://issues.apache.org/jira/browse/BEEHIVE-829 for details. Fixed Apache Beehive issue 834, "JDBC Control throws NPE after deserialization." See http://issues.apache.org/jira/browse/BEEHIVE-834 for details. Weblogic 9.0 Beehive v1m1 ships with XMLBeans v2. Changes were made to the JMS Control to fix some deserialization problems. |
|||
|
The application.xml file contained in weblogic-beehive-1.0.ear causes errors when parsed with the Xerces XML parser. The application.xml file contained in Remove the whitespaces at the beginning of the file: 1. Extract all files in |
|||
|
The build-control-module task references a CLASSPATH property that does not exist. If you use the build-control-module task and see the following error message during the build process:
you need to modify the Ant build script that calls it. Add the following line to the Ant build script:
after importing
|
|
On HP-UX, when an I/O operation on a file descriptor was cancelled, the socket was not getting properly cleaned and this caused a File Descriptor leak: this was leading to an OutofMemory exception. Contact BEA Customer Support for the one-off patch to fix this problem. |
|||
|
WebLogic Server does not enforce unique names for configured components. It is possible to configure two different components of the same type that have the same name. The Use unique names for components you create. Consider adopting a naming convention for new components. |
|||
|
If a client has installed a reverse-proxy server with IP mapping, an IP address and domain name may be externally exposed and discovered by someone outside the firewall. |
|||
|
Specifying a server start script used by Node Manager to start a server is supported only for the Java version of Node Manager, not the shell script version of Node Manager. |
|||
|
The Upgrade Wizard does not complete the database configuration when upgrading singleton service migration to server migration. Use the Administration Console to complete this configuration. See Server Migration. |
|||
|
A native library ( If you are on a platform that does not support |
|||
|
Changes to the configuration are protected by a lock based on the ID of the user that initiates the edit session. While one editor has the lock, other users are prevented from making modifications to the pending configuration. Because this lock is based on user ID, it is possible for two processes to modify the pending configuration at the same time if they are using the same user ID. Simultaneous edits with the same user ID is not recommended because it is extremely difficult to coordinate activities. A runtime exception occurs if two processes attempt to activate a pending configuration at the same time. If possible, avoid having two processes with the same username access an Administration Server. Use weblogic.Admin (deprecated) rather than WLST. Alternatively, when you use WLST, use the |
|||
|
When you configure cross-cluster session replication, setting the Session Flush Interval and Session Flush Session attributes to very low values can cause poor performance when the cluster attempts to send the replication data to an unresponsive remote cluster. |
|||
|
Large scale deployments on clusters with 10 or more nodes may cause out-of-memory errors. |
|||
|
When running on Solaris, the Node Manager logger intermittently throws an |
|||
|
The |
|||
|
Node Manager may not be able to find the |
|||
|
The value of the |
|||
|
jCom has not been updated to the most recent version and has not been tested for this Beta release. |
|||
|
The |
|||
|
After you migrate a JTA service, changes to the HostingServer available from the JTAMigratableTarget do not appear in the Administration Console. Because the changes do not appear, it is not obvious whether the migration succeeded or failed. |
|||
|
(HealthCheckPeriodsUntilFencing * HealthCheckIntervalMillis) < FencingGracePeriodMillis is enforced in the code. Currently, the following equation determines the length of the lease: (HealthCheckPeriodsUntilFencing * HealthCheckIntervalMillis * 2) + FencingGracePeriodMillis A server that crashes may not be restarted successfully by the NodeManager on the current machine because it thinks the lease is still active. A server that crashes is automatically migrated to a new machine where it will start successfully. So, the only negative effect is slightly increased downtime and a migration. The unnecessary migration can be avoided by increasing the |
|||
|
If you misconfigure a channel, your instance of WebLogic Server may become unstable. For example, if you accidently point to an incorrect IP address, you will see an error message in the server log, but you will be unable to access the Administration Console via the original port/protocol. The only way to regain access to the Administration Console is to restart all of the servers. |
|||
|
If you use a multi-threaded WebLogic Server 8.1 client (rather than a WebLogic Server 9.0 client) with a WebLogic Server 9.0 server , threads immediately fail with RJVM class cast exceptions. Retry a few times. The problem only seems to occur with a newly booted server. |
|
The |
|||
|
The
the tool attempts to activate a file specification named "true", because the |
|||
|
If you distribute an application to one or more Managed Servers that are unavailable (for example, because they are shut down or temporarily disconnected from the Administration Server), and the Administration Server is restarted, the unavailable server instances do not receive the distribute request when they become available. This problem does not occur with deployment, undeployment, or redeployment requests. |
|||
|
If you deploy an application to a cluster and one or more clustered servers are unavailable (for example, servers partitioned from the cluster due to a network outage), the deployment operation may appear to hang. In addition, the partitioned servers may not deploy the application even after they successfully rejoin the cluster. Reboot the partitioned servers after they rejoin the cluster. |
|||
|
ComponentMBean names generated during deployment are inconsistent between modules within an Enterprise application and stand-alone modules. Modules within an Enterprise applications receive a component name with a full file extension, while stand-alone modules receive a component name without a file extension. For example, if you deploy an Enterprise application containing an EJB named |
|||
|
This release of |
|||
|
The deployment API does not provide an implementation for the JSR-88 |
|||
|
A call to the JSR-88 |
|||
|
When you use the Administration Console to deploy multiple applications within a single change list, clicking Activate Changes deploys only the first application installed. The remaining applications are listed as "Distribute Initializing." To deploy the remaining applications, select the application names on the Deployments page and use the Start/Restart button. |
|||
|
The Builder and DDInit utilities, used to generate deployment descriptors, are not provided in this release. |
|||
|
If a configuration change is made to a DTD-based application, the change is persisted to the deployment plan, as it is for schema-based applications. However, the application does not pick up the new values from the deployment plan, either dynamically after a reboot of the server, or after a redeploy of the application. The Administration Console displays a message after a configuration save to indicate to the user that the persisted changes will not be used for DTD-based applications. The example applications installed in this Beta release use older, DTD-based WebLogic Server deployment descriptors. Configuring Applications for Production Deployment in Deploying Applications to WebLogic Server provides a link to an updated sample application that uses newer schema-based deployment descriptors, and that can be configured using a deployment plan. The Single Threaded Servlet Pool Size option on the Web Application configuration page in the Administration Console is deprecated and will be removed for the GA of this release. Currently, changing this option in the Administration Console has no effect on the Web application whether it is DTD or schema-based. |
|||
|
When deploying only one module of an application, WebLogic Server deploys the entire application (all modules) to all the targets on which any single module was deployed or targeted. Although not apparent in the Administration Console, the application is accessible through all of the available servers. |
|||
|
Using a common working directory in a Web application that is deployed on a cluster may result in corrupted or locked files. |
|||
|
For applications that have the Web Services module, WebLogic Server currently does not support side-by-side application deployment, which allows for the reliable introduction of new application functionality without disrupting the existing traffic to the deployed application. |
|||
|
If you try to deploy an application that references a missing library, the application will not start. |
|||
|
WebLogic Server does not detect the removal of application files when the server is not active, which may result in the domain tree not being in sync. Remove the autodeploy application while the server is in an active state. |
|||
|
If you pin an application to a particular server in a cluster and then deploy or redeploy that application to the cluster, the original target is left in the target list. Future attempts to undeploy the application from that single server will fail with an exception stating that the server cannot undeploy an application from a subset of the cluster. |
|||
|
When a plan is used to update a descriptor, the updated in-memory descriptor is not validated against any schema. |
|||
|
When deploying some services via an ant target, the XBus cluster DRT (part of remote.all) may fail on remote machines and throw the following error: The domain edit lock is owned by another session in exclusive mode in XBus cluster DRTs. |
|||
|
When deploying an application in a multi-machine environment, remote Managed Servers that are not targeted or do not belong to a cluster where some or all of the servers are targeted may get the following error message in the Managed Server logs:
Despite the above exception in the Managed Server logs, no adverse affect or incorrect behavior will occur because the Compatibility MBeans are not needed on these servers. |
|||
|
Killing the deploy command causes the activate operation to stay in preparing state for one hour. If you start a deployment operation via weblogic.Deployer and kill the command before the deployment actually starts, the underlying configuration manager activate operation may be left in the preparing state until the activate times out (one hour). If you attempt to perform another configuration manager save, undo, or activation operation during this time, you receive the following error:
Cancel the edit session, which resets the state for the activate operation that is incorrectly left in the preparing state. You can cancel the edit session in the Administration Console or with the WebLogic Scripting Tool (WLST). To cancel the edit session in the Administration Console, log into the Administration Console and then click "Lock and Edit". To cancel the edit session using WLST, use the |
|
Previous versions of WebLogic Server documentation describe Instead of using this non-public class, use JNDI to retrieve MBeanHome. See Determining the Active Domain and Servers in Programming WebLogic Server JMX Services. |