Age | Commit message (Collapse) | Author | Files | Lines |
|
- Introducing hazelcast for queue and progress map
- process batch of 100 at the time
- decreased module sync watchdog sleeptime to 5 seconds
- separate module sync tasks in new class and some other async preparations and easier testing
- tests for batching in module sync watchdog
- remove qualifiers annotation (support) where no longer needed
Issue-ID: CPS-1210
Issue-ID: CPS-1126
Signed-off-by: ToineSiebelink <toine.siebelink@est.tech>
Change-Id: I0a7d3755bf774e27c5688741bddb01f427d4a8a7
|
|
Implemented methods to perform a batch operation on updating datanodes
Refactored replace data node(s) tree methods to update data node(s) and descendants
Issue-ID: CPS-1203
Signed-off-by: lukegleeson <luke.gleeson@est.tech>
Change-Id: I365d657422b19c9ce384110c9a23d041eaed06f4
|
|
|
|
- update lombok config to handle Qualifier annotation
- update Semaphore config to use ConcurrentMap
- update SyncUtils to return a list of cm handles
- update DataSyncWatchdog and ModuleSyncWatchdog with Qualifier
- update DataSyncWatchdog to handle a list of cm handles
- Use get with xpath to check cm handle state
Issue-ID: CPS-1015
Change-Id: Icb39bd29f89e0020d49a1f8960476ffe81b12362
Signed-off-by: kissand <andras.zoltan.kiss@est.tech>
|
|
|
|
hasAllModules was returning no CMHandles when queried by itself, only when paired with another query condition e.g. hasAllProperties
This fix means endpoint works as expected.
Reverted null check
Issue-ID: CPS-977
Signed-off-by: lukegleeson <luke.gleeson@est.tech>
Change-Id: Idafad4c9877941d121f5cbb98617c56a35a1cc94
|
|
- seems like bug is not present on master
- modified test to ensure the behavior is tested now
- (intelliJ) cleaned up so unused stuff
Issue-ID: CPS-1212
Signed-off-by: ToineSiebelink <toine.siebelink@est.tech>
Change-Id: I6882df4277e50ef869c6a1e9b093ee83dbcdf78e
|
|
- Raise LCM Event when state transition happens.
- Raised events for ADVISED to READY, ADVISED to LOCKED, LOCKED to
ADVISED.
- Refactor existing code to raise the events correctly.
- Refactored existing test scenarios to comply with the code change.
Issue-ID: CPS-1034
Change-Id: Ie548e644f6133304d7fa36c892ca2bec7393c074
Signed-off-by: mpriyank <priyank.maheshwari@est.tech>
|
|
- State handler will now handle new structure of LcmEvent.
- We have 3 types of events i.e create, update and delete.
- Introduced the LcmEventCreatorHelper to delegate some of the
responsibility of event creation based on event type and the state.
- New tests and existing refactoring
- Code rebased
- Refactored name as per group code review
- Code rebase with other commits
- Used copy constructor for deep copy operation of NcmpServiceCmhandle
and CompositeState
- UPCOMING : Related user stories to trigger the event publishing using
state handler.
Issue-ID: CPS-1128
Change-Id: I94b5a87d37d6a174c017ee0aa37cd0f0f74ba084
Signed-off-by: mpriyank <priyank.maheshwari@est.tech>
|
|
|
|
|
|
|
|
An empty body for the CmHandles id-searches api returns an anchor instead of all cm-handles
Now returns all childnode ids for the dmi-registry
Issue-ID: CPS-1175
Signed-off-by: lukegleeson <luke.gleeson@est.tech>
Change-Id: I358970170a4eb7905e9646605f0f373fbcd78945
|
|
Added withCpsPath condition parameter
Validated to prevent misuse and blocking of querying using private properties
Updated OpenAPI with examples and links to documentation
Moved methods related to cmHandle querying using cps path from InventoryPersistence to CmHandleQueries
Renamed private method deleteSchemaSetAndListElementByCmHandleId to deleteCmHandleByCmHandleId
Issue-ID: CPS-977
Change-Id: I83827215b7e58de74f8f62cd0140516d217d93f1
Signed-off-by: lukegleeson <luke.gleeson@est.tech>
|
|
-Create API Which will enable/disable data sync enabled flag
-Default functionality of module sync watchdog is to set to false
-Remove global config param
-Will set initial sync state based on data sync enabled flag
-Throws an Exception if the same data sync enabled flag tries
to be set
-Throws Exception if state is not in READY
-Data Sync enabled must be true to complete data sync process
- Delete all resource data within fragment table related
to synced cm handle when data sync is set to false
Issue-ID: CPS-1133
Change-Id: Ib47bbd8293f083c1d705d91bd0def74e6a105c72
Signed-off-by: DylanB95EST <dylan.byrne@est.tech>
|
|
- use semaphore map in ModuleSyncWatchdog
- increase test timeout, because it needs more time for hazelcast
initialization
Issue-ID: CPS-1015
Change-Id: I71feed8fbbd047af9fabba29a5f762a1f17a1c78
Signed-off-by: kissand <andras.zoltan.kiss@est.tech>
|
|
- Created a dedicated threadpool for scheduler.
- Tuned async threadpool of notification executor from setting RejectedExecutionHandler and application.yml.
Issue-ID: CPS-1126
Signed-off-by: sourabh_sourabh <sourabh.sourabh@est.tech>
Change-Id: I2afe3c76c1aec78751777df0d2f08ddb8dcee102
|
|
- refactored method "parseAndRemoveCmHandlesInDmiRegistration"
- added method to send events for DELETING and DELETED
- introduced private method in state handler
- updated unit tests for registration spec
- added unit test to cover new private method in state handler
Issue-ID: CPS-1003
Signed-off-by: emaclee <lee.anjella.macabuhay@est.tech>
Change-Id: Ifc1d1e94b7103d85b6352b196edf7075257c211a
Signed-off-by: sourabh_sourabh <sourabh.sourabh@est.tech>
|
|
-Old schema deleted and comments from tests are removed
-EventDateTimeFormatter utility added
-From LcmEventsCreator.populateLcmEvent the NcmpServiceCmHandle param
removed and tests are fixed accordingly
-Changes made in test of the LcmEventsCreator for payload testing
to make it more clear
Issue-ID: CPS-1147
Change-Id: Id87c81f2755a6617a0fc68ba92a3d0e3170fd20f
Signed-off-by: bmiklos <miklos.baranyak@est.tech>
|
|
- Refactored code to reflect the LcmEvents scope now.
- Test cases updated and dependencies reviewed.
- You can find few occurences of NcmpEvent for now but that will be
cleaned up once we implement CPS-1104 which is for LcmEvent schema
- UPCOMING: LcmEvent schema and then the code to publish the events on
cmHandleStates
- LATER : LcmEvents based on public properties update.
Issue-ID: CPS-1137
Change-Id: I9f395ed733b4028e706205894c36a38412e3452e
Signed-off-by: mpriyank <priyank.maheshwari@est.tech>
|
|
- State handler taking care of ADVISED, READY and LOCKED state
transition at the moment.
- For now I have not removed the actual code but eventually state
handler will take care of persisting the state and publishing of the
events
- Rebased code to add the global parameter related to dataSyncCache
- UPCOMING : The classes will be prefixed/renamed with LCM once we have the
LcmEvent created with new schema
Issue-ID: CPS-1118
Change-Id: Ic45d95169eb0c06cfb35c907d34380dbcbf2da11
Signed-off-by: mpriyank <priyank.maheshwari@est.tech>
|
|
|
|
- Define the initial Data Sync Cache enabled through
configuration parameter
- Set the data sync enabled flag based on this
- And in turn define the initial sync state of the
Data Sync
Issue-ID: CPS-1119
Change-Id: I43bf03c79481291bf47c9b672f7bf408d789df61
Signed-off-by: DylanB95EST <dylan.byrne@est.tech>
|
|
|
|
Issue-ID: CPS-475
Change-Id: Id24f2b0c6fe95ca9c7559a8dbd160c4566ddc32f
Signed-off-by: bmiklos <miklos.baranyak@est.tech>
|
|
|
|
|
|
- Introduce the hazelcast dependency and check how the builds react to
it.
- Data Sync and Module Sync Map config.
- Added basic test scenarios.
- Updating the artifact-id from hazelcast-all to hazelcast-spring and
version is 4.2.5. (5.1.2 is latest but has known vulnerabilities)
- Refactored the code as per the onsite review session
- Upcoming : Actual sub-task which will use the distributed maps in
respective use cases.
Issue-ID: CPS-1050
Change-Id: I60314c4ff95c267b186b20862aef93a5c389a971
Signed-off-by: mpriyank <priyank.maheshwari@est.tech>
|
|
Issue-ID: CPS-1117
Signed-off-by: emaclee <lee.anjella.macabuhay@est.tech>
Change-Id: If6a0c620970a2a34a601267c6610f85e4bc07f60
|
|
-created new method in inventory persistence service for getting module
references
Issue-ID: CPS-1117
Signed-off-by: emaclee <lee.anjella.macabuhay@est.tech>
Change-Id: Ic1f3c180a62a7211e19982d8f3570829db58370e
|
|
Refactored LOCKED_MISBEHAVING -> LOCKED_MODULE_SYNC_FAILED
CompositeStateMapper will change internal reason LOCKED_MODULE_SYNC_FAILED to external reason LOCKED_MISBEHAVING for client payloads
Changed openapi description of lock-reason to reflect only enum currently available LOCKED_MISBEHAVING
Issue-ID: CPS-1099
Signed-off-by: lukegleeson <luke.gleeson@est.tech>
Change-Id: I9cda45f6c30b94684ee1c8ad0c49e35a3a824d52
|
|
query"
|
|
query
Issue-ID: CPS-1085
Change-Id: Ifac152f606b25b7a7d143315dba6d73698ef914e
Signed-off-by: kissand <andras.zoltan.kiss@est.tech>
|
|
|
|
Allows the Resync of an already synced CmHandle
Currently: Advised -(create schemaset)-> Ready -(manual write)-> Advised -> Locked as schemaset for cmhandle already exists
With this: Advised -(create schemaset)-> Ready -(manual write)-> Advised -(delete schemaset, create schemaset)-> Ready
Included some logging
Renamed ModuleSyncSpec -> ModuleSyncWatchdogSpec to match class
Issue-ID: CPS-1045
Signed-off-by: lukegleeson <luke.gleeson@est.tech>
Change-Id: I408fbea698b7926dbf5d0cddc74acf1b00235b1f
|
|
- part of this commit includes renaming the enum SyncState to
DataStoreSyncState
Issue-ID: CPS-1064
Signed-off-by: emaclee <lee.anjella.macabuhay@est.tech>
Change-Id: I6bf419141a1b33f09871946445cdfff422c8c354
|
|
|
|
|
|
|
|
-NOTE: Build will fail until docker-compose version issues on build
server are fixed
--Ticket raised https://jira.linuxfoundation.org/plugins/servlet/theme/portal/2/IT-24219
-added flag for async
-added response if async is triggered without being enabled & associated test
-modified to use one global flag for notifications
Issue-ID: CPS-1088
Signed-off-by: JosephKeenan <joseph.keenan@est.tech>
Change-Id: If9d988b4dcb71bf37c1b1bf9464090782708ffc2
|
|
- do NOT use cpsPath query for gettign individual CM Handles
- change order of queries for more efficency (property query first)
Issue-ID: CPS-1085
Change-Id: Ia3bdf4f0d245a2b128557352685d2aa17c1f939c
Signed-off-by: kissand <andras.zoltan.kiss@est.tech>
Signed-off-by: ToineSiebelink <toine.siebelink@est.tech>
|
|
Added new get cm-handle state endpoint
Refactored RestOutputCmHandleState to CmHandleCompositeState
Created new RestOutputCmHandleCompositeState OpenApi object
^This is done so that we get '"state: {" at the start of JSON response
Refactored RestOutputCmHandleStateMapper to CmHandleStateMapper
Added more detailed composite state to get cmHandleDetails endpoint tests
Rebased code
Code rebased on top of 129658: Unable to change state from LOCKED to ADVISED | https://gerrit.onap.org/r/c/cps/+/129658 which fixes output error
Issue-ID: CPS-1019
Signed-off-by: mpriyank <priyank.maheshwari@est.tech>
Signed-off-by: lukegleeson <luke.gleeson@est.tech>
Change-Id: I361117c98c256a4aa578c13d21926bc6d7876a15
|
|
Retry algorithm for module-sync based on last failure
Poll Lock Reason, check if lock has surpassed minimum time
based on last update time and lock reason
Issue-ID: CPS-1076
Change-Id: Ifbbabd2b403f88f1bbe3fae3f125b1e9cb2091aa
Signed-off-by: DylanB95EST <dylan.byrne@est.tech>
|
|
-Refactor Cm Handle Registration to deal with null
pointer during Module Sync Retry
-Add Last Update Time to cm handle registration
-Add Fetch Descendants Option as paramater
for get misbehaving locked cm handles
-Fixes issue when state that goes to advised
after retry mechanism won't be retried again
Issue-ID: CPS-1097
Change-Id: Iffe1cef3479a796ea5c78b293a0bd24a86d13efd
Signed-off-by: DylanB95EST <dylan.byrne@est.tech>
Signed-off-by: sourabh_sourabh <sourabh.sourabh@est.tech>
|
|
|
|
- Remove operation field from NcmpEvent schema as it is not required
- Code refactoring corresponding to removal of operation field
- Fixed test cases around it
- UPCOMING: Some sort of state machine to handle the event publishing
Issue-ID: CPS-1100
Change-Id: Ib89b86bd1408de2808b01d37a48ee44a2601f83d
Signed-off-by: mpriyank <priyank.maheshwari@est.tech>
|
|
Issue-ID: CPS-1101
Signed-off-by: lukegleeson <luke.gleeson@est.tech>
Change-Id: I527b47cd98150a753c0d7ba89332a52b414b6834
|
|
- Get all the Cm Handles state in READY and Operational datastores sync state in UNSYNCHRONIZED
- Get a random Cm Handle
- Get the first resource data from the node
- Save the data in Cps Db
- Update the Operational datastores sync state to SYNCHRONIZED
Issue-ID: CPS-1052
Issue-ID: CPS-1053
Issue-ID: CPS-1054
Change-Id: I9a20391ef30e6d56c4d789a92b8bf42cd3756c62
Signed-off-by: Lathish <lathishbabu.ganesan@est.tech>
Signed-off-by: sourabh_sourabh <sourabh.sourabh@est.tech>
|
|
|
|
|