Age | Commit message (Collapse) | Author | Files | Lines |
|
|
|
Issue-ID: CPS-2510
Change-Id: I4093459c824c202dc8dec4b869e338d4b80fbac8
Signed-off-by: seanbeirne <sean.beirne@est.tech>
|
|
Issue-ID: CPS-2478
Change-Id: I0aa3804225f053ff99176056fe71322a23cff31c
Signed-off-by: ToineSiebelink <toine.siebelink@est.tech>
|
|
|
|
Issue-ID: CPS-2510
Change-Id: I7068d0de45cc2c2e5d8815f0e78c260369f3551e
Signed-off-by: seanbeirne <sean.beirne@est.tech>
Signed-off-by: ToineSiebelink <toine.siebelink@est.tech>
|
|
The endpoints for fetching modules and module definitions
previously returned empty result if a CM handle was not
found or not ready. Recent changes for alternate ID have
changed the behaviour so it returns an error instead.
This restores original behaviour by swallowing the exception.
Change-Id: I63091d04d171e6f4ac6ad1c44b359b9a0c4b8f65
Issue-ID: CPS-2509
Signed-off-by: danielhanrahan <daniel.hanrahan@est.tech>
|
|
Issue-ID: CPS-2386
Change-Id: I0d1f09cff202e055be255b365fc2271a5ee43b37
Signed-off-by: seanbeirne <sean.beirne@est.tech>
|
|
This reverts commit 7e6fee9321e4f13cc4a0d81f83fc3693309cb5d7.
Reason for revert: This patch is causing blocking issues at ESH see CPS-2509
Issue-ID: CPS-2509
Change-Id: Ie1bd26ccca4abcac09faf35e2b97491dab32b14d
Signed-off-by: ToineSiebelink <toine.siebelink@est.tech>
|
|
|
|
Issue-ID: CPS-2386
Change-Id: Ib4f37858f5f107e350a13e4bf7ed93493864ce60
Signed-off-by: seanbeirne <sean.beirne@est.tech>
|
|
- Add (micrometer) instrumentation to expose inefficiencies
- Add test config for micrometer
- Add setup methods in base to create many cm handles
- Set module sync parallelism to 2 for testing
- Add clean up methods for hazelcast related tests
- added test to show inefficiencies
- POC 1 use hazelcast set to prevent multiple threads working on same ModuleSetTag
- POC 2 'cache' module set tags per thread to prevent DB looks ups
- Main inefficiency left: create schemaset for EACH cm Handled even if same tag. No easy PoC...
Change-Id: Idf46b44c475a24727dd7084bb613459f4c29be55
Signed-off-by: ToineSiebelink <toine.siebelink@est.tech>
|
|
|
|
|
|
Issue-ID: CPS-2481
Change-Id: I8a200495be1afb9a4f256127f886bed78c77f653
Signed-off-by: seanbeirne <sean.beirne@est.tech>
|
|
NCMP should cut off the part after the # of an alternateid
and use only the first part in CPS match check, but send the complete request to the DMI plugin.
Issue-ID: CPS-2485
Change-Id: Icc1442f2be9545036619043692c3559ffadecb0d
Signed-off-by: egernug <gerard.nugent@est.tech>
|
|
|
|
|
|
|
|
dmi-data-service-name for writeDataJob"
|
|
for writeDataJob
-Change implementation to convert datanode to yang model
-Change to resolve service name consistent with other implementations
Issue-ID: CPS-2455
Change-Id: I9cbe494b8c37eb963c42abee1ac5f04a3f695d07
Signed-off-by: egernug <gerard.nugent@est.tech>
|
|
Issue-ID: CPS-2274
Change-Id: I75163bc7943fca5cbbe8c24601a2dfc184a9e6cc
Signed-off-by: seanbeirne <sean.beirne@est.tech>
|
|
- Fixed checkstyle and license validation.
- Updated checkstyle configuration (`onap-java-style.xml` and `check-license.xml`) to ensure compatibility with Maven Checkstyle Plugin version 3.3.1
- Removed obsolete properties causing build failures.
- Improved code style rules to meet ONAP Java guidelines.
- Vulnerabilities scan are now configured on parent pom.xml
- Fixed all code style.
Issue-ID: CPS-2470
Change-Id: I1e72af604415aa7587cbe1dff471ed0323e29aa3
Signed-off-by: sourabh_sourabh <sourabh.sourabh@est.tech>
|
|
- improve readablity of ModuleSyncTasks
Issue-ID: CPS-2474
Change-Id: Ib0342e8fad00b431a293c5095fe54f4917f7a51b
Signed-off-by: halil.cakal <halil.cakal@est.tech>
|
|
- incoming Cm Avc Event from DMI Plugin is consumed and forwarded to
target topic
- the key from source topic to be used in the target topic while
forwarding
- with same key the ordering of the message will be preserved
- NOTE: the RTD related changes will be a separate patchset
Issue-ID: CPS-2436
Change-Id: Ie692663706b378022ec0d621d92ca5054bad8d1b
Signed-off-by: mpriyank <priyank.maheshwari@est.tech>
|
|
This commit moves the state update code into the 'finally'
block, so that state updates are always sent for the handles
that finished syncing, even if an exception occurred.
This commit also fixes a typo in cmHandleStatePerCmHandle.
Issue-ID: CPS-2474
Signed-off-by: danielhanrahan <daniel.hanrahan@est.tech>
Change-Id: Ic25608a5772c0873fe0d0f8aca8d70bcda51d94a
|
|
Currently, module sync will output messages such as:
"cm-handle-1 is now in READY state"
before the state change is persisted to the DB. If an exception
occurs, the new state may not be persisted, which makes debugging
hard due to misleading log entries. This commit moves the logging to
the LCM state handler immediately after the change is persisted.
- Move logging of state changes to LcmEventsCmHandleStateHandler.
- Remove unused code in LcmEventsCmHandleStateHandler.
- Add tests of state change logging.
- Update the main test showing the bug to instead show expected
behaviour, but disable test for now using @Ignore.
Issue-ID: CPS-2474
Signed-off-by: danielhanrahan <daniel.hanrahan@est.tech>
Change-Id: I58646af6df95e07e69ca525744713a7c01c25e41
|
|
|
|
|
|
Issue-ID: CPS-2402
Issue-ID: CPS-2383
Change-Id: I5dd3132b70b401bf2f50f77c2c131d5d34aa0c0a
Signed-off-by: seanbeirne <sean.beirne@est.tech>
|
|
It was determined that one CSIT is intermittently failing due
to a CM-handle being deleted while module sync is in progress,
which causes the whole batch operation to fail. Even CM-handles
that did sync will not go into READY state, despite the logs
saying otherwise. This commit reproduces the issue in a unit test,
and prevents the issue in the CSIT by changing test order. Also,
errors during module sync tasks are reported at ERROR level.
(The actual bug fix will be addressed in another patch.)
Issue-ID: CPS-2474
Signed-off-by: danielhanrahan <daniel.hanrahan@est.tech>
Change-Id: I7e0d617cbd48d8fd1fad036079fbd876ee21d8a8
|
|
- Add tests for multi threaded scenarios around module sync
- Disabled ModuleSyncWatchdog timer using long delay and interval
- Call Module Sync method as needed for more control
(sometimes it needs to be triggered twice like retry use cases as designed)
- Improve NCMP performance test setup (consistent naming etc.)
- Rename some production code method names to better reflect functionality
- Disabled intermittent failing test for create cm handle as it is not asserting the correct message
- Improved Code Coverage ModuleSyncWatchdog
Issue-ID: CPS-2462
Change-Id: Ia907af77d2037309f1bbb73ea671679b788bab9e
Signed-off-by: ToineSiebelink <toine.siebelink@est.tech>
|
|
- Consolidated multiple `if` statements to enhance readability and maintainability.
- Replaced traditional `instanceof` checks followed by casting with the new pattern matching syntax for improved type safety and clarity.
- Removed unnecessary casts and streamlined code in `addYangLeafList` and `addDataNodeFromNormalizedNode` methods.
Issue-ID: CPS-1957
Change-Id: I46696dc184b6b3c3d0fe3a6bd12267a4e0d76193
Signed-off-by: sourabh_sourabh <sourabh.sourabh@est.tech>
|
|
- update spotbugs to 4.8.6
- remove old findbugs dependency (spotbugs replaces findbugs)
- remove an unused SupressFBWarnings
Issue-ID: CPS-2458
Signed-off-by: danielhanrahan <daniel.hanrahan@est.tech>
Change-Id: If9d5571ae4293a2653a117a988414af2452647c0
|
|
Queue Management
- Enhanced logging for populateWorkQueueIfNeeded and resetPreviouslyLockedCmHandles
methods, improving clarity and error handling.
- Improved readability and maintainability of the locking mechanism with
Hazelcast's FencedLock.
- Optimized error handling in catch blocks, logging detailed exception messages and stack
traces for better troubleshooting.
- Refined lock acquisition and release flow, with clear log messages for both successful
and failed lock operations, ensuring safe handling of Hazelcast distributed locks.
Issue-ID: CPS-2403
Change-Id: Ie089f36a817d4965782235b51ee987ef054516b1
Signed-off-by: sourabh_sourabh <sourabh.sourabh@est.tech>
|
|
- removed old entries of data structures which are removed now.
- updated the documentation to correctly reflect our infrastructure
startegy related to hazelcast.
- fixed few NB comments on the merged patch.
Issue-ID: CPS-2408
Change-Id: Id7ce4ac695ad94742490ac5d2be89cd23ffda3e5
Signed-off-by: mpriyank <priyank.maheshwari@est.tech>
|
|
- updated all the cache(distributed datastructures) to use the same instance config now.
- there will be just one instance per JVM now
- Better cleanup of hz instance in the testware
- Refactored the testware to verify the configs
- Expected Impact on lowering the memory usage, less number of TCP
communications between members
- NOTE: we need to do a full regression as the changes impacts all the
use cases which involves cache
Issue-ID: CPS-2408
Change-Id: I7564992a9990f44ef3defb4f50cb7d094cad7b92
Signed-off-by: mpriyank <priyank.maheshwari@est.tech>
|
|
(fighting between IntelliJ and Checkstyle best practices)
- non-2xx responses are processed using web client exceptions
- handle unknown host exception
- upgraded spotbugs (checkstyle and related mvn plugin)
- fixed some small spotbugs due to upgrade
- added commented instructions in docker compose to enable debugging
- added some environment variables for policy executor configuration
- extract out Sleeper in stub service to achieve 100% coverage
- added cause to Policy Executor exceptions where applicable
- ignored (new) spotbug rule about catch NPE because of issue in 3pp
- ignored (new) spotbug rule about \n in string due to multiline string block
Issue-ID: CPS-2412
Change-Id: I6835a73320c436cbeea12cc7a06f15899eec7bf1
Signed-off-by: ToineSiebelink <toine.siebelink@est.tech>
|
|
- return NONE for get effective trust level api if the trustlevel caches
empty (restart case)
Issue-ID: CPS-2409
Change-Id: I72a755ca6ba5d4a78f5458a235c0d1f43aaa8d53
Signed-off-by: halil.cakal <halil.cakal@est.tech>
|
|
- configurable default answer
- apply default answer upon non 2xx response
- delayed default webclient read timeout
- add custom timeout method with original read timeout in seconds
- apply default answer upon timeout
- add integration test with short timeout error scenario
Issue-ID: CPS-2412
Change-Id: I62527a27e426c2f01fda2182ebd2513242c29ac1
Signed-off-by: ToineSiebelink <toine.siebelink@est.tech>
|
|
This patch removes the AnchorDataCache from CPS, which is used for
prefix resolution in get/query operations.
As such, Hazelcast is no longer a dependency of CPS, only NCMP.
- Changed PrefixResolver to be more efficient.
- Removed AnchorDataCache and associated classes.
- Moved HazelcastCacheConfig to NCMP.
- Removed Hazelcast dependency from cps-service/pom.xml
This shows good performance improvements in some APIs such as v2 GET
which is nearly 2x faster (also 5x faster including base patch).
Issue-ID: CPS-2417
Signed-off-by: danielhanrahan <daniel.hanrahan@est.tech>
Change-Id: I24768469f24e90b70f7a6187faa4f5b3d75777d2
|
|
This reverts commit https://gerrit.onap.org/r/c/cps/+/138853
Reason for revert: Commit caused performance issues in CM handle search
Manual rebase was required to integrate changes for alternate-id
support made since original commit was merged.
Issue-ID: CPS-2375
Signed-off-by: danielhanrahan <daniel.hanrahan@est.tech>
Change-Id: I5ef743ea053589fa17c10587317a3534c2714c4b
|
|
watchdog poll
- Increased watchdog frequency for locked cm handle.
- Removed retry backoff algorithm for locked cm handle.
Issue-ID: CPS-2395
Change-Id: I54d0ec8f9de53a7d181639c14aaaa93736f03e19
Signed-off-by: sourabh_sourabh <sourabh.sourabh@est.tech>
|
|
|
|
- fixed import order and moved package structure
Issue-ID: CPS-2293
Change-Id: Ie2f9f057f261577054530feee7480850ba4b41e1
Signed-off-by: leventecsanyi <levente.csanyi@est.tech>
|
|
|
|
- Remove servers from yaml.
- Add prefix to path in yaml.
- Remove basepath prefix from Controller Impl.
- Added policy server port to 8093.
- Exposed docker env. vars for policy services.
Issue-ID: CPS-2291
Change-Id: Idcc9e23f9d63bad2480537dc8d39523fe82b4b83
Signed-off-by: ToineSiebelink <toine.siebelink@est.tech>
Signed-off-by: sourabh_sourabh <sourabh.sourabh@est.tech>
|
|
Issue-ID: CPS-89
Change-Id: Ia6e3bb26ffb433f3610f53f408fad12489f2e3b3
Signed-off-by: halil.cakal <halil.cakal@est.tech>
|
|
CpsDataPersistenceService is an internal interface of CPS, part of
the SPI, intended only to be used by CPS. Instead, NCMP should use
the public API of CPS, i.e. CpsDataService and CpsQueryService.
This is an architectural issue, but it will affect upcoming changes,
where changes to the SPI implementation should not affect public API.
Issue-ID: CPS-2398
Signed-off-by: danielhanrahan <daniel.hanrahan@est.tech>
Change-Id: I89339e1a3566a25e33168c24c8b8401ca52224e9
|
|
Some special case code to disable ancestor-axis was added for
CM-handle search (see CPS-2308). It is now relocated to NCMP.
This makes other needed improvements of Cps Path Parser easier.
- Move special case code into NCMP
- Add integration test to ensure CM-handle search works
Issue-ID: CPS-2365
Signed-off-by: danielhanrahan <daniel.hanrahan@est.tech>
Change-Id: I168d6156be559166f115aa42e21cd987d98b7d41
|
|
|