SR-B44278 · Issue 302830
MT guide updated for rules blocking/restriction
Resolved in Pega Version 7.3
The MultiTenant guide has been updated to clarify what causes rules to be blocked or restricted and the possible consequences of unrestricting rules.
SR-B48780 · Issue 305215
Oracle/Unix documentation updated with JDBC instructions for urandom instead of random
Resolved in Pega Version 7.3
The Install/Upgrade guides have been updated to reflect that recent changes to Oracle have created the requirement of forcing the JDBC driver to use /dev/urandom instead of /dev/random to improve security as well as connection open and close performance. The guide now recommends customers running Oracle on Unix do one of the following: 1. Update $JAVA_HOME/jre/lib/security/java.security to use securerandom.source=file:/dev/./urandom 2. Set system property -Djava.security.egd=file:///dev/urandom
SR-B8847 · Issue 278000
FCM/FCMR documentation updated
Resolved in Pega Version 7.3
The documentation on the PDN for installing and configuring FCM 7.12 has been updated for FCM and FCMR.
SR-B8898 · Issue 279449
Reference to outdated property removed
Resolved in Pega Version 7.3
When building a system with WebSphere 8.5.5, DB2 10.5fp7 and Java 8, creating the environment and testing the connection on its datasource produced the warning: "DSRA8200W: DataSource Configuration: DSRA8020E: Warning: The property 'progressiveLocators' does not exist.". This was traced to a code fragment left after that property became unsupported which has now been removed.
SR-B64134 · Issue 313542
Workaround added for Oracle 4000 character column value max
Resolved in Pega Version 7.3.1
Due to an Oracle limitation, RD on a clob column fails if the column value is of length greater than 4000. To work around this, support has been added to retrieve data from a clob column of length greater than 4000 characters using report definition based on DASS setting "reporting/retrieveFullClobContent" defined on "Pega-RULES".
SR-B51989 · Issue 310393
TimeStamp column includes milliseconds in extract
Resolved in Pega Version 7.3.1
Extraction to database output format was not showing the millisecond part of the datetime column. This was a missed use case for BIX when the support of Datetime property mapping to timestamp column was introduced in Oracle, and has been fixed.
SR-B64125 · Issue 313855
Fixed SLA calculations with custom classes in Oracle
Resolved in Pega Version 7.3.1
For work classes, columns for pySLAGoal and pySLADeadLine are of type TIMESTAMP in Oracle DB. The Pega getContentItemsPastSLA activity uses the columns to find date differences with SYSDATE and then used a subtraction operation between columns of type TIMESTAMP and SYSDATE to calculate the SLA. However, an exception resulted when custom classes were used. To correct this, getContentItemsPastSLA has been updated to cast the columns of type TIMESTAMP to DATE type.
SR-B66297 · Issue 314861
Instance insertion fixed for PostgreSQL partitioning
Resolved in Pega Version 7.3.1
Obj-Save was failing when trying to insert an instance in a partitioned PostgreSQL table with 'only if new' enabled. This record insertion failed because of a system check that is present when inserting an instance into the child table, and the issue has been resolved by reworking the code that was checking (assuming) if the insert was successful based on the returned inserted count.
SR-B66297 · Issue 299892
Instance insertion fixed for PostgreSQL partitioning
Resolved in Pega Version 7.3.1
Obj-Save was failing when trying to insert an instance in a partitioned PostgreSQL table with 'only if new' enabled. This record insertion failed because of a system check that is present when inserting an instance into the child table, and the issue has been resolved by reworking the code that was checking (assuming) if the insert was successful based on the returned inserted count.
SR-B51825 · Issue 310042
Added handling for WO lock when OAuth2 token expires
Resolved in Pega Version 7.3.1
The expiration of the OAuth2 token committed the WOs in context and caused all WOs to lose their locks and eventually end up in the SLA broken queue. To fix this, code has been added to perform the commit on a new database thread so the commit operation will be independent from the actual thread context.