Clone Tools
  • last updated a few minutes ago
Constraints
Constraints: committers
 
Constraints: files
Constraints: dates
[WFCORE-4748] Allow the MODULE_OPTS to be set and if an -javaagent: is found in the options automatically add jboss-modules as an agent to the VM.

https://issues.redhat.com/browse/WFCORE-4748

  1. … 35 more files in changeset.
[WFCORE-4415] Upgrading JBoss Modules from 1.9.0.Final to 1.9.1.Final This update brings in the following fixes: * [MODULES-387] Add JMX operation to get a class location * [MODULES-386] Fixing Javadoc build issues on JDK11 * [MODULES-375] Check if the artifact was successfully resolved before attempting to create the loader. * [MODULES-372] provide our own "java.se" module on JDK 9+

  1. … 7 more files in changeset.
[WFCORE-4183] Fix problem of incorrectly including runtime-only attributes in read-resource by dropping support for invalid-for-years-now inclusion of attributes without a ManagementResourceRegistration entry.

  1. … 5 more files in changeset.
[WFCORE-4183] Fix problem of incorrectly including runtime-only attributes in read-resource by dropping support for invalid-for-years-now inclusion of attributes without a ManagementResourceRegistration entry.

  1. … 5 more files in changeset.
[WFCORE-3039] Capability requirement can be lost if two attributes on same resource reference the same capability

  1. … 11 more files in changeset.
[WFCORE-3039] Capability requirement can be lost if two attributes on same resource reference the same capability

  1. … 11 more files in changeset.
[WFCORE-4000] Refactoring org.jboss.as.process.jdk.JdkType -> org.jboss.as.host.controller.jvm.JvmType

  1. … 9 more files in changeset.
[WFCORE-4000] Prevent domain server VM validation breaking things.

1) Defer VM validation until ManagedServerBootCommandFactory.getServerLaunchCommand, to ensure failures are handled consistently with other kinds of server launch problems.

2) Refactor other uses of ManagedServerBootCommandFactory so instead of using getServerLaunchCommand they can call use-case specific API, allowing the factory to skip inappropriate validation, reserving it for the launch process itself.

  1. … 8 more files in changeset.
[WFCORE-4001] Eliminating --illegal-access=permit option from WildFly.

  1. … 14 more files in changeset.
[WFCORE-4000] Move JdkType from "version" module to "process-controller" module.

  1. … 11 more files in changeset.
[WFCORE-4039] Fixing JvmOptionsBuilderFactory JDK decision logic.

  1. … 3 more files in changeset.
[WFCORE-4000][WFCORE-4001] Automagically propagating necessary modular JDK options.

  1. … 13 more files in changeset.
WFCORE-3944 NPE in SyncServerStateOperationHandler when group is not present on host

  1. … 1 more file in changeset.
Allow mock OperationContext to supply requisite capability registrations.

  1. … 1 more file in changeset.
[WFCORE-3683]: Enable a resource to define requirements without binding them to an attribute. Passing the resource registration address to build the requirement capability pattern.

  1. … 23 more files in changeset.
[WFCORE-3465] Get rid of extraneous uses of ModelNode.resolve(). Related, get rid of impls of the unused ParameterValidator.validateResolvedParameter and implement a default method that throws UOE

  1. … 36 more files in changeset.
[WFCORE-3423] Refactoring - no functional change. Eliminating usage of MSC service listeners in test suite.

  1. … 7 more files in changeset.
[WFCORE-3139] Make it possible to get all the attribute info in one call; use that in ValidateModelStepHandler

  1. … 11 more files in changeset.
WFCORE-1374 Update build for jdk9 b175+

  1. … 39 more files in changeset.
[WFCORE-2964] Use the default-server-logging.properties file on initial boot of domain servers. Ignore the logging.configuration system property which should always be set to the incorrect file by the domain startup scripts.

  1. … 2 more files in changeset.
WFCORE-2784 Host controller should handle jdk9 specific jvm parameters

  1. … 5 more files in changeset.
[WFCORE-2829] Provide the ProcessType via the ManagementResourceRegistration

  1. … 9 more files in changeset.
WFCORE-1981 remove deprecated API

  1. … 7 more files in changeset.
[WFCORE-2707] Clean up the socket-binding-group and interface resource defs and handlers [WFCORE-2706] Expose the SocketBindingManager as a capability

  1. … 31 more files in changeset.
WFCORE-2202 - minor improvements to multiple dynamic parts api

  1. … 7 more files in changeset.
[WFCORE-1337] Expose paths and the PathManager as capabilities

  1. … 13 more files in changeset.
[WFCORE-1987] - add warning capability to core.

  1. … 8 more files in changeset.
[WFCORE-350]: Domain reflects jboss.server.xy properties

Filtering host properties starting with jboss.server. when creating the server launch command line.

    • -0
    • +135
    ./as/host/controller/ManagedServerBootCmdFactoryTestCase.java
  1. … 1 more file in changeset.
[WFCORE-2094] Domain server notifies HC upon controller instability

This commit introduces a notification mechanism via which a listener for "controller instability" notifications can be created and wired into the ModelController.

In a domain server, such a listener is registered. In this case it's an additional function of HostControllerClient, which responds to notifications by asynchronously sending a message back to the managing HC.

Currently the managing HC simply logs a WARN when receiving such a message. In the future it could also do something else with the data, e.g. expose it over the management API for use by the admin console.

Notifications are currently sent in two cases:

1) An OperationContext is interrupted or times out waiting for MSC stability during operation finalization. This means MSC was not stable at the time of finalization and the state of the system is no longer knowable.

2) An Error is thrown during operation execution. Again in such a situation the state of the system is no longer knowable.

In the latter case, there is no guarantee that the listener will be able to send the message to the HC. For example if the Error is an OOME sending a message may not be possible. So this is not a foolproof means of ensuring the HC is aware of server instability. Possible future work to cover that kind of scenario could be to have the HC send some sort of health check request to the server when it detects that an operation has timed out.

This also doesn't send a notification if the server fails to shut down properly, e.g. from problems with the MSC ServiceContainer not stopping. That would be detected by the bootstrap shutdown hook, which has no relationship to this logic.

  1. … 14 more files in changeset.
[WFCORE-2094] Domain server notifies HC upon controller instability

This commit introduces a notification mechanism via which a listener for "controller instability" notifications can be created and wired into the ModelController.

In a domain server, such a listener is registered. In this case it's an additional function of HostControllerClient, which responds to notifications by asynchronously sending a message back to the managing HC.

Currently the managing HC simply logs a WARN when receiving such a message. In the future it could also do something else with the data, e.g. expose it over the management API for use by the admin console.

Notifications are currently sent in two cases:

1) An OperationContext is interrupted or times out waiting for MSC stability during operation finalization. This means MSC was not stable at the time of finalization and the state of the system is no longer knowable.

2) An Error is thrown during operation execution. Again in such a situation the state of the system is no longer knowable.

In the latter case, there is no guarantee that the listener will be able to send the message to the HC. For example if the Error is an OOME sending a message may not be possible. So this is not a foolproof means of ensuring the HC is aware of server instability. Possible future work to cover that kind of scenario could be to have the HC send some sort of health check request to the server when it detects that an operation has timed out.

This also doesn't send a notification if the server fails to shut down properly, e.g. from problems with the MSC ServiceContainer not stopping. That would be detected by the bootstrap shutdown hook, which has no relationship to this logic.

  1. … 14 more files in changeset.