INC-132088 · Issue 603086
Updated Tracer for cluster events
Resolved in Pega Version 8.4.4
When running the tracer from the Admin Studio Listener management landing page, Tracer was always fetching data from one node only even though events were generated on another node. This was due to the specific node where the Listener was running being passed to the concern activity while launching the tracer; the Tracer Client node did not have data to check if the event was being populated on another node, and the result was to always fetch data from one node only even though events were generated on another node. To resolve this, the web node ID where the event is running will be passed.
INC-134188 · Issue 576731
XML Parse handles dates with time zone correctly
Resolved in Pega Version 8.4.4
Due to a difference in data mapping behavior for Date values, the Parse XML new assembler implementation was incorrectly applying the DateTime type of processing for Date inputs when using timezones and silently ignored the exceptions reported. This has been resolved.
INC-137775 · Issue 583783
Improved resolved rules cache
Resolved in Pega Version 8.4.4
When Rule resolution iterated over a candidate list to fetch a candidate, performance issues were seen on very large sites. To resolve this, an enhancement has been added that will cache the resolved virtual table entries to optimize performance in high demand use cases like DSM.
INC-137882 · Issue 599140
Hazelcast will be used to determine node presence
Resolved in Pega Version 8.4.4
When calling logoff from the end user portal, sporadic PresenceStandardLifeCycleListener exceptions were seen during high load periods. This was traced to changes made in the Service Discovery behavior: when there is an issue with the node and heartbeats are missed, the node goes to Safe Mode. This caused the topology to change and trigger clearing off presence sessions, but the node still received service requests and this issue was seen. To resolve this, an updated has been made so the system will rely on the Hazelcast listener instead of service discovery to determine whether a node has left.
INC-138253 · Issue 597569
Handling added to ensure proper Data-Admin-DB-Table name records
Resolved in Pega Version 8.4.4
As part of property optimization, After exposing the property "SolutionName" using the wizard, the value was not visible in a new column for existing records after the completion of the column population job. This was traced to an issue with DBSchemaDDLExecutor expecting lowercase Data-Admin-DB-Table names, and the wizard has been updated to ensure Data-Admin-DB-Table name records are all lowercase.
INC-138743 · Issue 602446
Updated LegacyPageAdapter with forward links
Resolved in Pega Version 8.4.4
After upgrade, attempting to submit an assignment from an existing case type did not work, and the message "There was an error processing the 'ValidateTechnicalData' step [(no task)]: java.lang.ClassCastException: com.pega.platform.clipboard.adapters.internal.LegacyPageAdapter cannot be cast to com.pega.pegarules.data.internal.clipboard.ClipboardPageImpl" appeared on screen. Investigation showed that a reference forward link was unexpectedly referring to a LegacyPageAdapter. This has been resolved by implementing get/setNamedPageReferenceForwardLinks in LegacyPageAdapter so LegacyPageAdapter can be cast to a ClipboardPageAPI.
INC-138840 · Issue 599980
Data-Summary reports default will not use the rule resolved setting
Resolved in Pega Version 8.4.4
An intermittent issue was seen where a query could become stuck for ~2 hours in the database. This has been resolved by setting the default for Data-Summary reports to not use the rule resolved setting.
INC-139297 · Issue 601422
JSON content type update
Resolved in Pega Version 8.4.4
An update has been made to ensure the content_type is set to application/json for JSON response.
INC-141838 · Issue 601680
Locking and classmap improvements
Resolved in Pega Version 8.4.4
Threads were getting stuck inside classMap and tableInformationMap, causing nodes to go down. This was traced to issues with classmap performance and locking, and has been resolved by replacing the lock library with a newer version and refactoring the tableInformationMapImpl class to use a concurrent hashmap instead.
INC-142188 · Issue 600331
Resolved parent lock sending child item to broken queue
Resolved in Pega Version 8.4.4
If the ServiceLevelEvents agent tried to process the queue item for a child case but encountered a lock error because the parent case was locked, the queue item was immediately going to the Broken-Process status instead of re-queueing the item to try again in 15 minutes. If the property was set to 'Copy data from a data page', the item was re-queued. To resolve this, the unwind and forward chain calls have been reversed. A dynamic system setting is available to reset the change.