These bugs have been fixed in JBoss Operations Network 3.1.1. For a complete list of all issues addressed in JBoss Operations Network 3.1.1, search Bugzilla for bugs with a target of JON 3.1.1.
If resource plug-ins were deleted when resources of that type were in the inventory, then an agent running in the background on a system was stopped and had errors when restarting.
If arguments used with the exec command within the JBoss ON CLI had quotation marks, the arguments were not properly parsed and the script execution failed.
Attempting to run the retrieveBackingContent method for a new resource could fail for up to 10 minutes after the new content resource was imported into the inventory.
There was a problem decoding obfuscated passwords used to connect to content providers like the Customer Portal. This caused connection attempts to fail, even with valid credentials.
CVE-2012-0022. The JBoss Web version used by JBoss ON used an inefficient approach for handling parameters, which allowed remote attackers to cause a denial of service attack by increasing CPU consumption through a request that contained many parameters and parameter values.
Having more than 1000 agents in the inventory for a single server could cause errors in the server log with the SQL exception ORA-01795: maximum number of expressions in a list is 1000.
Groups could display or include uncommitted resources in group counts as members. This included recursive groups, which could include uncommitted child resources of explicit members.
If a compatible group were created with a description longer than 100 characters, the creation failed with the error value too long for type character.
Only one trait can be used in a dynamic group expressions. However, if multiple traits were set, the UI returned a stack trace that the group members could not be calculated. Now, the UI properly returns an error that only one trait is allowed and does not throw a stack trace.
Empty groups (with an UNKNOWN group availability status) were incorrectly returned in searches for groups with up, down, or disabled availability states.
Empty groups were not listed in the Groups page, even if they were successfully created and visible in other areas of JBoss ON (such as the bundle deployment wizard).
Using the Recipe option to upload a recipe file separately than the bundle archive files required that the XML in the recipe be well-formed and that all tags were explicitly closed (not the abbreviated format <tag />).
If a new version of a bundle contained configuration properties or tokens and the original version of the bundle did not contain any configuration properties, then attempting to deploy the later version would fail. Likewise, if a new configuration property were added to a configuration file, then attempting to deploy the later bundle version with the new parameter would fail.
If a bundle version was deleted, it was not properly removed from the listed bundle versions. This caused bundle deployments of newer bundle versions to throw exceptions and fail.
New information has been added to debug messages for metric collections to help identify if a particular metric were delaying metric collections for an agent or a resource.
When manually importing a content-backed resource, the resource ID was not properly set, which prevented the resource from being properly added to the inventory.
If request method used with the SNMP modules for Apache was null or empty, then the module would segfault when it attempted to use the uninitialized memory. Now, the empty request is treated as an undefined protocol, and the module returns a message that the request method (null) is unsupported.
A new JMS destination can only be created by giving the required name and JNDI name, with the other values supplied by default settings. However, the defaults were not included in the create operation, so attempting to create a new JMS destination using the defaults failed.
EJB2 MDB resources were always shown as down (unavailable), even if they were started and running. The availability component was not able to set the run state for the EJB2 MDB resource, so the plug-in was never able to move the resource out of an unknown or down state.
If a JBoss EAP server became unstable due to OutOfMemoryError condition or other problem, the JBoss AS 5 plug-in was prevented from connecting to the server. In that case, the resource was reported as DOWN, even if it were really up. The agent then would not invoke a shutdown operation for the "down" resource, even if the resource were actually still running.
When a new child was created, a discovery scan did not run immediately. This meant that the new child was undiscovered until the next scheduled discovery scan, which could be several minutes or even hours. Now, a discovery scan runs as soon as a new child is created successfully, so that it is discovered immediately.
An edit icon appeared next to module names in the Configuration > Current > Extensions area of the EAP 6 server configuration. These properties are read-only and cannot, in fact, be edited.
The JBossWeb connector configuration has a section for SSL parameters, and those settings were only valid if SSL is enabled (the secure option is selected). However, these fields were still active even if SSL were not enabled. Also, disabling and re-enabling SSL blanked all of the previous SSL configuration.
The UI allowed a user to attempt to create a second JVM definition as a child to a managed server or server group. If a JVM definition already existed for a managed server or server group, then an error was returned when the user attempted to save the second JVM definition. Now, the user is prevented from creating a second definition.
All child elements for an embedded HornetQ resource required a JNDI name in the configuration, but the UI did not force the JNDI name to be given when creating the children. This meant that the creation operation failed.
Attempting to change the configuration for a security domain failed because the security domain properties were set in the UI as a simple list rather than a map. Because the value should have been a map, the new simple property value that was submitted did not validate.
Some HornetQ resource properties need to have a value of null of the property is blank, but the configuration was being sent to the server as empty. Using an empty value caused any configuration change to fail because the values couldn't be validated.
The memory pools name property for the garbage-collector resource is read-only (with only one potential value) but was not marked as read-only in resource configuration. This meant that a user could attempt to change the configuration for the memory pools name property, but the configuration change would fail.
Several EAP 6 properties were set to accept only long values and would reject configuration changes with other values, but the actual EAP 6 properties (such as retry-interval-multiplier for the HornetQ subsystem) accept decimal values. If a decimal value were read from the EAP 6 configuration, then the user was (incorrectly) prompted to change the decimal value.
Attempting to change the configuration for a security domain failed with the error Received 500 (Internal Server Error) response to Operation, and the configuration remained the same. However, the configuration change was incorrectly reported as being successful.
Managed servers were not discovered when a domain or host controller's name was determined at runtime instead of being specified in the host.xml configuration file. The agent log has the message Requested management resource not found: host=master, meaning it was looking for a resource with the name master rather than the host configuration name given at runtime, $(hostname).
The agent incorrectly attempted to deploy resource to an EAP 6 standalone server in the deployment/ directory rather than the real deployments/ directory.
When creating a new cache container in the UI, all of the templates (except for the default template) had improper mode properties set, which resulted in invalid configuration when creating the cache container through the UI.
The reload operation for the EAP 6 server failed with the error Credentials for plugin to connect to AS7 management interface are invalid - update Connection Settings with valid credentials, even if the credentials were correct. The problem was that the operation actually succeeded, but that the JBoss ON agent may check for the operation success while the server was still in the process of restarting, so it incorrectly assumed that the operation had failed.
The profile resource incorrectly reported some configuration properties as empty, even if the default value appeared in the managed server. When a required value was incorrectly reported as missing, opening the configuration in the JBoss ON UI required the user to change the configuration.
This is actually an underlying EAP 6 problem, not a problem with the EAP 6 plug-in. The properties which were not being properly detected have been removed from the configuration to prevent spurious or unnecessary configuration changes.
Both the JBoss Data Grid and the JBoss EAP 6 plug-ins can discover and create Infinispan cache resources. If both JDG and EAP 6 were installed on a system, and JBoss ON was using both the EAP 6 and JDG plug-ins, then both plug-ins would attempt to create or discover the Infinispan cache resource, regardless of the actual server parent of the cache resource. This led to unpredictable discovery behavior, including the resource not being discovered at all.
EAP 6 can allow configuration either for static connectors or for the discovery group name, but they cannot be defined simultaneously. Attempting to change the configuration for one area, such as the static connectors, while the other area was defined resulted in configuration update failures.
This is actually an underlying EAP 6 problem, not a problem with the EAP 6 plug-in. The allow-direct-connections-only property for the static connector has been removed from the configuration to prevent invalid configuration changes.
If the connector property was unset, then it was not possible to create a connection factory or pooled connection factory resource. The server incorrectly sent a null key for the map value of the connector property. The configuration has been updated to send an actual null rather than a map with null:null set as a key value.
When editing certain HornetQ child resources, the configuration could be edited in a way that caused the EAP 6 server to crash and not be able to restart. Several properties (static connectors, connectors, and discovery group name) have been changed to read-only on created resources and cannot be changed once a resource is created. This prevents any changes to the resource which create an invalid configuration file.
If there were more than 200 children for a resource, not all of the children are displayed in the resource tree or refreshing the tree might have displayed a different subset of child resources.
In Internet Explorer 8 only, attempting to upload content such as an EAR file failed at several points in the UI, including creating a new child in the inventory, uploading a bundle, and uploading an alert script.
Some information about the server cannot be changed during upgrade. Some fields, such as the server name, were editable in the upgrade form which should not have been. Changing these values from the original settings caused NPEs in the server logs and the upgrade to fail.
When the JBoss ON server is upgraded from 3.0 to 3.1, then the UI could return a 404 error when opening the installer page to re-install the server. The browser incorrectly cached a redirect to the old UI, server.example.com:7080/coregui. Because the coregui/ path did not exist yet, because the new server had not yet been configured, this redirect returned a 404 error.