INC-
184212 · Issue 677286
Updated AgentName handling for QueueItemID
Resolved in Pega Version 8.6.4
A report was showing as scheduled but no mail was received when it was supposed to run. Investigation showed this was due to the reports being corrupted, leading to the flow skipping the necessary Queue-For-Agent method. While there was a workaround of doing a "Save As" to create a new version of the report, this has been addressed by setting the agentName before saving to the database to handle missing agentName cases and ensure pyAgentname is always populated when pzQueueItemID is not empty.
INC-195586 · Issue 698861
Updated access group handling for CurrentWorkPool property
Resolved in Pega Version 8.6.4
After update, the pxThread.pxCurrentWorkPool property was not properly populated in App Requestors when the activity was called from Rest service. This was caused by a difference in the authentication check after a security modification, and has been resolved.
INC-197479 · Issue 695025
ClusterAndDBCleaner updated to with with Oracle query limits
Resolved in Pega Version 8.6.4
The pzClusterAndDBCleaner job scheduler was not able to cleanup data in pr_op_data session table due to the delete query formed to clean up this table throwing "ora-01795 maximum number of expressions in a list is 1000 oracle 19c" exception. This has been resolved by splitting requestor IDs into batches of 1000.
INC-197964 · Issue 707875
Engine updated to default to hard references in cache
Resolved in Pega Version 8.6.4
A scheduled campaign executing at midnight was failing with a null pointer error message. Rerunning the same campaign during the day did not encounter any issues. Investigation showed there was a race condition that occurred under heavy load and parallel executions which caused soft references to be garbage collected while a thread was accessing the same soft references. This has been resolved by updating the following DSS so the system will default to using hard references: Ruleset: Pega-Engine Purpose: prconfig/cache/enablesoftreferences/default Value: false
INC-199076 · Issue 703301
Revalidate of class rules successful after major version skimming
Resolved in Pega Version 8.6.4
When performing a major ruleset version skim and revalidating all rules, most class rules were failing during the revalidation. This was a gap in how Ruleset List and Major Version Skim work together. When calling getRuleSetPrerequisites(“rulesetA:01-01-01”), the API uses uses the current application context to return a ruleset list from an application which owns rulesetA and its compatible version, but if no owning application was found from the current application context, a list of rulesetA and PegaRULES:XX-XX-XX was returned. This caused a conflict between what the system was trying to validate and the ruleset versions returned. To resolve this, an update has been made which will exclude the version number for this comparison if no owning application is found.
INC-199192 · Issue 689038
Check added to ensure StepPage generation for Top keyword
Resolved in Pega Version 8.6.4
A ClassCastException was thrown when top page is used, indicating "com.pega.pegarules.data.internal.clipboard.ClipboardPropertyImpl cannot be cast to com.pega.pegarules.pub.clipboard.ClipboardPage". This has been resolved by adding a check which will generate a myStepPage for Keyword "Top" as necessary.
INC-199341 · Issue 694714
Content-type values made consistent
Resolved in Pega Version 8.6.4
The content-type for service APIs was returned as 'text' for 401 errors when using a custom authentication type instead of the parent-type/child-type format such as text/plain, text/html, application/json, etc. This has been corrected.
INC-202183 · Issue 699683
ClusterAndDBCleaner updated to with with Oracle query limits
Resolved in Pega Version 8.6.4
The pzClusterAndDBCleaner job scheduler was not able to cleanup data in pr_op_data session table due to the delete query formed to clean up this table throwing "ora-01795 maximum number of expressions in a list is 1000 oracle 19c" exception. This has been resolved by splitting requestor IDs into batches of 1000.
INC-204998 · Issue 705630
Data page definition reload made more robust
Resolved in Pega Version 8.6.4
The data page was intermittently being removed from the cache. This was traced to the use of a thread from ThreadContainer to reload the data page definition: a null thread could be returned if the request came from a master agent, which would then cause the definition reload to fail. This has been resolved by updating the system to use the current thread in context to reload the data page definition.
INC-207009 · Issue 701554
Explicit expiration added to avoid searching for expired requestor
Resolved in Pega Version 8.6.4
A login page was taking long time to display. This was traced to pre-authentication cookie in the browser pointing to the requestor object on the server which triggered a lookup across the entire cluster of servers to find the requestor. This was not only taking time, but the attempt to find the requestor in the cluster would always fail to return results as the requestor was not passivated but instead removed after two minutes. To resolve this, an expiration has been added to the Pega-RULES cookie when the value is pre-authenticated. The time to expire is derived based on the short-lived requestor time for unauthenticated requestors + 1 minute, and will be 2 minutes by default. This will avoid searching for a requestor across all nodes in cluster when the requestor has already timed out and been destroyed by server.