Age | Commit message (Collapse) | Author | Files | Lines |
|
|
|
Also updated licenses.
Replace tabs with spaces in new junit method.
Updated licenses again.
Change-Id: Iebff19353c052e59bfd69da1677e9a05c44592a0
Issue-ID: POLICY-869
Signed-off-by: Jim Hahn <jrh3@att.com>
|
|
|
|
When a new operation is started it will throw an IllegalArgumentException
if a property is missing. Modified the rules to catch this
exception, indicate the failure, and move on to the next operation.
Modified code to catch generic exception.
Change-Id: Ie8c24371eb84ee55ff0a0a5fcacda3670483efc5
Issue-ID: POLICY-869
Signed-off-by: Jim Hahn <jrh3@att.com>
|
|
Because the lock owner string is constructed, using just the target type
and request ID, lock requests made by different ONSETs running on
different PDP-Ds may both succeed. Added the hostname to the owner
string so that locks for ONSETs for the same request ID will be rejected
if the request is made on different PDP-Ds.
Change-Id: I1d45194cbb1c5b1a1f6000477fe51879a2fecc19
Issue-ID: POLICY-872
Signed-off-by: Jim Hahn <jrh3@att.com>
|
|
If rules are deleted, then no rule is left to retract the
associated Params object. However, we can create more general
rules that will fire whenever rules are added, thus causing
clean-up of old Params at that time.
Implemented the above approach.
Change-Id: I207d493cc6924a4b34a8d0fa4915ed499ebc4a6e
Issue-ID: POLICY-872
Signed-off-by: Jim Hahn <jrh3@att.com>
|
|
When rules for a closed loop are deleted, associated working memory
elements may be left around.
Modified the code to retract the various elements.
Re-added code to unlock when retracting TargetLock.
Add junit tests for new methods.
Forgot the junit test for request id.
Change-Id: Iec2805adbdb147a643f6163badee05ea353ada88
Issue-ID: POLICY-872
Signed-off-by: Jim Hahn <jrh3@att.com>
|
|
ControlLoopEventManager.unlock only returns the TargetLock if
the PolicyGuard.unlock operation is successful. However, if
the lock has already expired, then PolicyGuard.unlock will return
false, and ControlLoopEventManager.unlock will then return null.
As a result, the rules do not retract the TargetLock. (The rules
do, however, retract any left-over TargetLock objects during the
clean-up phase. Nevertheless, the earlier rules should not have
to depend on the clean-up rules to do that.)
Modified the code to address this issue.
Change-Id: I2e6099a4e3511053d2e6e1373cae4b480798d1b6
Issue-ID: POLICY-872
Signed-off-by: Jim Hahn <jrh3@att.com>
|
|
Modified the local policy guard locking facility to add
a time parameter.
Modified the control loop event manager to extend the lock
when lockCurrentOperation() is re-invoked.
Modified the rules to retract the lock if the lock request
was denied.
Reorder assertions in junit test.
Change-Id: Ic9b77acbb4881a5a516f30eb56664bad1a5c4d7e
Issue-ID: POLICY-872
Signed-off-by: Jim Hahn <jrh3@att.com>
|
|
|
|
The kie-maven-plugin causes a lifecycle management
error in Eclipse (not in normal build). A profile
is added to the POM to remove this eclipse
error.
Change-Id: I6f5e5bec5e1e388edb083a910b9d1472af889a47
Issue-ID: POLICY-716
Signed-off-by: liamfallon <liam.fallon@ericsson.com>
|
|
Snapshot updated for Casablanca.
Issue-ID: POLICY-875
Change-Id: I0307282c4eabdb9ccd03282f918f4cd727ac4f86
Signed-off-by: liamfallon <liam.fallon@ericsson.com>
|
|
Modified code to skip building credentials if user name
is empty, in addition to case where user name is null.
Change-Id: I5b53d4f97263df98cfe082f44a4e6af64b1e488b
Issue-ID: POLICY-868
Signed-off-by: Jim Hahn <jrh3@att.com>
|
|
|
|
- reduce memory footprint
- avoid verbose logging from cache when
the object is not in the cache to
invalidate it.
Change-Id: I819cfbcfe32f5afdd238397c79ccd5cebd4534b8
Issue-ID: POLICY-852
Signed-off-by: Jorge Hernandez <jh1730@att.com>
|
|
These reviews must be merged in order because of layering.
This review should be merged THIRD.
Issue-ID: POLICY-844
Change-Id: I61d128f3244cd420abab15ab70c5c0e9d9afa625
Signed-off-by: liamfallon <liam.fallon@ericsson.com>
|
|
The code was passing two different sets of credentials to the REST
manager, and the wrong one was taking precedence. Modified the code
to only pass the correct set of credentials.
Updated license date.
Change-Id: I13e2db93a8eebbb0cd77fdfaca80125cd28a2553
Issue-ID: POLICY-800
Signed-off-by: Jim Hahn <jrh3@att.com>
|
|
|
|
These reviews must be merged in order because of layering.
This review should be merged FOURTH
PLD added version.properties
Change-Id: If7af7b3ae37166e88af92c0d1f1c558e5b60ab8e
Issue-ID: POLICY-798
Signed-off-by: liamfallon <liam.fallon@ericsson.com>
Signed-off-by: Pamela Dragosh <pdragosh@research.att.com>
|
|
Change-Id: I3b04d6be4939cc2e652a676e7c2e4cb647f1bfec
Issue-ID: POLICY-796
Signed-off-by: Jim Hahn <jrh3@att.com>
|
|
Change-Id: I0040a3be47d56808e2cda8a5ab2d3c1aa8e9fbb9
Issue-ID: POLICY-796
Signed-off-by: Jim Hahn <jrh3@att.com>
|
|
The requestId keeps creeping into this code. For the instance
Id appending, we need a better strategy that conforms with robot
test scripts in order to do multiple scale ups.
Issue-ID: POLICY-793
Change-Id: Iab5d7608d5ee049cb6dd0508108bf77c2982db47
Signed-off-by: Pamela Dragosh <pdragosh@research.att.com>
|
|
It seems this field is missing possibly from refactoring code.
Issue-ID: POLICY-792
Change-Id: I6d13feb12a383d32dce3bbba431e32ad0b9d1e19
Signed-off-by: Pamela Dragosh <pdragosh@research.att.com>
|
|
this allows to change the dcase onset processing
group behavior across pdp-ds.
if installation sets DCAE_CONSUMER_GROUP, all
PDP-Ds will share the same consumer group, therefore
will have exclussive processing of onsets. Otherwise
if installation unsets it (blank), all pdp-d's will
process all onsets being placed in dcae topic on
dmaap bus
Change-Id: I4cc63005402ce66c09617b52dfad9a6094cbcb46
Issue-ID: POLICY-766
Signed-off-by: Jorge Hernandez <jh1730@att.com>
|
|
- Changing abatement parameter to true for vCPE in create-cl-amsterdam script
- Enabling abatement test case jMeter Test plan
Change-Id: I01487666a807941dfec729c68bc31d3cedf29967
Issue-ID: POLICY-525
Signed-off-by: eikrwaq <waqas.ikram@ericsson.com>
|
|
AAI simulator fixes for multi onset messages in parallel
Change-Id: Ia32dacf6caaded7fc45232ba4be8f49f3ccfd614
Issue-ID: POLICY-525
Signed-off-by: eikrwaq <waqas.ikram@ericsson.com>
|
|
|
|
Modified PolicyGuard to use the locking API from policy-core instead
of its own internal map.
Initialized PolicyGuard.factory.
Fixed bug in eventmanager junit test uncovered due to changes to
PolicyGuard.
Change-Id: I853ee5f146f3bde16b3f6e65bc188ca7c0cc4f73
Issue-ID: POLICY-577
Signed-off-by: Jim Hahn <jrh3@att.com>
|
|
Turns out that the getAttributesTest() method depends on
testGetCountFromDb() being executed first, but junit does
not guarantee this. Modified the code so that it is
independent of the order in which the tests are executed.
Change-Id: Ie081061e4e83895f4ac89242cf00d843b6e069d4
Issue-ID: POLICY-765
Signed-off-by: Jim Hahn <jrh3@att.com>
|
|
In the rules, the $manager was missing the condition to match the
request id of the manager with the request id of the event. As a
result, the completed operation was being applied to an unrelated
manager.
Change-Id: Icb63c0f151788e7f96450e26853270a7368dfcd8
Issue-ID: POLICY-763
Signed-off-by: Jim Hahn <jrh3@att.com>
|
|
Modified event manager code to not query AAI if there's already
a response from a previous query.
Modified code to store AAI properties in local variables instead
of static to prevent any chance of a race condition in a multi-threaded
situation.
Updated junit tests to use a new manager for each subsequent
query.
Update license.
Change variable name, in junit test, from onsetEvent to event,
because it can be any type of event.
Included fix to prevent initial AAI query if AAI data was already
available in the initial onset event.
Change-Id: Idf3e15ea8c5e297f22f23570c22fd837b72ba200
Issue-ID: POLICY-754
Signed-off-by: Jim Hahn <jrh3@att.com>
|
|
Per this link: http://www.baeldung.com/httpclient-4-basic-authentication
apache HttpClient does not send the credentials by default,
but instead waits for the server to issue a challenge before
sending them. Modified RESTManager to construct and send the
Authorization header always instead of waiting for the
challenge.
Remove unused imports.
Also add Authorization header for POST request.
Change error message to match original error message when
username is null.
Preempt sonar issue.
Allow null user name, which indicates that the Authorization header
should not be sent. Note: this only impacts the RESTManager class;
invoking classes (e.g., AaiManager) still enforce that the user name
exists in the properties.
Fix license dates in modified test code.
Change-Id: I3cb26e76562db746939631437775727809553390
Issue-ID: POLICY-754
Signed-off-by: Jim Hahn <jrh3@att.com>
|
|
I removed a couple of empty README and added licenses.
Issue-ID: POLICY-741
Change-Id: I10aad3536d02b738a1dc606f7311de6b3a2777cd
Signed-off-by: Pamela Dragosh <pdragosh@research.att.com>
|
|
Since we are not using LDAP PIP we can get rid of this jar
that has security issues.
Issue-ID: POLICY-722
Change-Id: I93feacc8733a834866476db75933d8b2cf08c212
Signed-off-by: Pamela Dragosh <pdragosh@research.att.com>
|
|
Issue-ID: POLICY-736
Change-Id: I2bc455510e925764d18a1d8e96e072bdadb251c6
Signed-off-by: Pamela Dragosh <pdragosh@research.att.com>
|
|
The "requestID" field was changed to "requestId" which brakes
the use cases since this does not conform to the events
DCAE will be sending. To avoid further issues with changed
variable names, all onset fields now have a serialized name
that will be used when serializing/deserializing with gson.
Issue-ID: POLICY-681
Change-Id: I731b0c83a9b6c9bbaae0312f76382d7f4644db8c
Signed-off-by: Daniel Cruz <dc443y@att.com>
|
|
Change-Id: I5305e71bb0cc844dd37ff6420cd30012f98c88ed
Issue-ID: POLICY-720
Signed-off-by: Jorge Hernandez <jh1730@att.com>
|
|
The following changes aim to have a quicker start of the drools container:
1. For both amsterdam (and the experimental beijing) controllers
dependencies are pre-installed, so the loading of 3rd party
dependencies are faster.
2. Further enhancements in installation.
3. Make sure that the naming of the generated control loop
artifacts have a "control loop" label associated with it,
so in the future, if other applications are added (non
control loop related) do not conflict and is clear.
Change-Id: Iecb84d186fcc34069aa5c4a175a8a4521b38499d
Issue-ID: POLICY-534
Signed-off-by: Jorge Hernandez <jh1730@att.com>
|
|
we are not using this code yet, but let's remove
jackson-databind. The inclusion of drools-pdp adds
in the latest jackson-databind.
Issue-ID: POLICY-504
Change-Id: I73a1939eb3827a7ef06c40e11764944b3d80685b
Signed-off-by: Pamela Dragosh <pdragosh@research.att.com>
|
|
Removed checkstyle warnings from:
policy/drools-applications/controlloop/common/model-impl/appc
policy/drools-applications/controlloop/common/model-impl/appclcm
policy/drools-applications/controlloop/common/model-impl/events
including changes to template files due to renamed methods
Issue-ID: POLICY-705
Change-Id: Ieb93baf43268aa608f204eef1a610354f2adec32
Signed-off-by: mmis <michael.morris@ericsson.com>
|
|
|
|
|
|
Change-Id: I3b411ee2855a1ffe94ffc67fccfb1e475335d036
Issue-ID: POLICY-533
Signed-off-by: Jorge Hernandez <jh1730@att.com>
|
|
|
|
The package path of LoggerUtil changed as below:
From:
import org.onap.policy.drools.utils.LoggerUtil;
To:
import org.onap.policy.drools.utils.logging.LoggerUtil;
Change-Id: I8187d97de5eeab92268baab155fb4de37f4df3ad
Issue-ID: POLICY-448
Signed-off-by: liamfallon <liam.fallon@ericsson.com>
|
|
|
|
The create-cl-beijing script will now allow the user to create a rules
jar that does not need to be expanded and also generate test files that
can be used to insert facts for all the supported beijing use cases. The
yaml for each policy is provisioned by the user in case custom yaml is
desired.
In addition to this, a new script is included that will extract the
needed data from the controller properties file to insert facts for all
the supported use cases. This script can be used to insert
ControlLoopParams facts after the create-cl-beijing script is used to
generate the rules artifact. No provisioning necessary other than
specifying where the controller properties file is located.
Issue-ID: POLICY-692
Change-Id: I42f0a08fba133ca36fb1be588a720e4f9598d79f
Signed-off-by: Daniel Cruz <dc443y@att.com>
|
|
Removed checkstyle warnings from
policy/drools-applications/controlloop/common/model-impl/aai
including renaming classes:
AAI* -> Aai*
PNF* -> Pnf*
and renaming some methods in those classes
Issue-ID: POLICY-705
Change-Id: I2d59a668728aa58a2c9fbe78c44e924e6cfb4531
Signed-off-by: mmis <michael.morris@ericsson.com>
|
|
|
|
The timeout implementation assumes that a finished request is a HTTP-200
together with either a request state of COMPLETE or FAILED in the embedded
request in the response in the JSON. This is the same as what is done earlier
in this class and in, for example, VFC.
Unit test for timeout has also been added.
Change-Id: I637dbecb8d230b8060f7ced76d92d11ec482503e
Issue-ID: POLICY-448
Signed-off-by: liamfallon <liam.fallon@ericsson.com>
|