SR-D898 · Issue 435997
Pega0069 alert accuracy resolved for UHC FinishAssignment
Resolved in Pega Version 8.2.2
Pega0069 alerts were not raised for the CS commit and confirm harnesses, and the reported total duration was not always accurate. This has been resolved by capturing document busy/idle for CS harnesses.
SR-D9132 · Issue 461283
Performance improvements for EventStoreWriter class
Resolved in Pega Version 8.2.2
The cache protocol has been modified In order to improve performance for the EventStoreWriter class.
SR-D9511 · Issue 465301
pyPreTriage and pyPostTriage flows now default to Draft Mode Off
Resolved in Pega Version 8.2.2
The out-of-the-box pyPreTriage and pyPostTriage flows had draft mode turned on. Due to this, Email Triage cases in production systems were sent into the problem flow. This has been corrected by turning off draft mode for pyPreTriage and pyPostTriage flows.
SR-D11382 · Issue 482269
QP DF issues resolved for timeouts and search re-indexing
Resolved in Pega Version 8.2.2
The re-index option was not available to manually trigger Search re-indexing. Investigation showed that this was related to a recent change used to hide the reindex button if the batch index queue processor was in progress. At the root of the issue, delayed items were not properly pushed to Kafka when they were eligible for processing if they were already queued when the QP DF was in a failed state or when the QP DF was not yet created. When restarted, the QP DF was only processing the messages on the Kafka topic from the end of the offset and missing the queued messages. To resolve this, the QP DFs have been modified to start processing the messages from the beginning of the topic. In addition, the DelayedQueueProcessorSchedule Job scheduler, which pushes items to Kafka, was not being executed due to a Hazelcast timeout exception caused by an inability to access the distributed map. Investigation showed this exception was related to a cluster split brain. To correct this, in cases where a distributed map is not accessible, the run will be rescheduled. As a result, there is now more robust handling for any timeout exceptions, all messages will be processed as expected, and the re-index button will be available when needed to manually trigger the process.
SR-C96394 · Issue 438671
Report data query updated to better handle exceptionally large queues
Resolved in Pega Version 8.2.2
If the database queue was exceptionally large (more than 3.5 million items), the report data query aggregating the results was timing out and showing 0 items queued on the management landing page. To resolve this, the summary view has been changed to rely on the data pages of the detailed view instead of aggregating the results by firing another query with RD.
SR-D8467 · Issue 456412
Resolved false alarm exception for Cloud BIX extraction
Resolved in Pega Version 8.2.2
BIX Extraction on Cloud was intermittently throwing an exception even though the BIX Extraction was successful. This false alarm was triggered by inconsistencies in an API within the BIX extraction module used to determine if files existed on a Cloud S3 location post-extraction, and has been resolved by removing the file check.
SR-C89954 · Issue 431931
Rule-Connect-SOAP error fixed for WAS/JBOSS environments
Resolved in Pega Version 8.2.2
After upgrade, SOAP webservice was failing with the error “Couldn't create OMElement from string value for the activity InvokeAxis2 in class Rule-Connect-SOAP". This occurred on sites using a WebSphere/JBoss environment, and was traced to an issue where the WAS classloader was loading its default jar at the same time the Pega classloader was loading the same jar into JVM. This has been resolved.
SR-C93572 · Issue 440079
Search initialization logic updated to ensure consistent node member information
Resolved in Pega Version 8.2.2
The search function was intermittently failing across nodes after restart. Investigation showed that this was due to inconsistent results from search initialization cluster logic using Hazelcast APIs to tell whether or not a given node was part of a Pega cluster. To resolve this, the logic has been updated to rely on ES APIs instead of Hazelcast cluster membership to determine offline nodes.
SR-D9242 · Issue 483173
Security improved for DownloadFromServiceExport
Resolved in Pega Version 8.2.2
In order to improve security, the activity “DownloadFromServiceExport” now requires authentication and explicitly blocks attempts to traverse up directories.
SR-D7427 · Issue 444636
Stack overflow error resolved for pxObjClass returned as declarative target
Resolved in Pega Version 8.2.2
After installation, logging in for the first time as [email protected] with the temporary password was generating an exception. Investigation showed that the system was attempting to obtain the class name of page containing a declarative target and ended up in recursion due to pxObjClass, which was read while getting class name, being returned as a declarative target from the conclusion cache. To avoid this, the code has been modified to ensure that pxObjClass can never be a declarative target.