SR-B73213 · Issue 322353
CMIS-GetPropertiesResponse property mapping fixed
Resolved in Pega Version 7.3.1
The CMIS properties' data was improperly mapped to a "pyProperty" Page List property which was not part of the CMIS-GetPropertiesResponse data model, rather than mapping the data into the proper "pyProperty" Page List properties in the CMIS-GetRepositoriesResponse data model such as pyPropertyBoolean, pyPropertyInteger. This was caused by a defect in the Pega engine's CMISConnector module caused by changes made to simplify and consolidate the "map Properties" logic. The CMISConnector module has been repaired so that it correctly derives data type when giving a pointer to a PropertyDataObject:
SR-B55054 · Issue 317660
Fixed Safari refocus on section refresh
Resolved in Pega Version 7.3.1
Safari was setting the focus to document.body after a section refresh inside a details portion, causing qualifying questions to disappear. The issue was found to be that the section displaying the questions was running under a separate thread. As long as the refresh happened within that context, the behavior was as expected. However, immediately clicking on an answer for the question without first clicking elsewhere within the window caused it to run under the claims context and not render properly due to the focus shift. To correct this, modifications have been added to avoid resetting the execution thread conditionally for Safari when the blur is triggered on document.body.
SR-B72892 · Issue 323517
Error message display fixed for RG read-only field
Resolved in Pega Version 7.3.1
Incorrect behaviors were seen with displaying errors when a field in a repeat grid table was configured as "Read-only (always)" and "Show validation messages in read-only mode" was on. In Google Chrome, the UI form did not contain the corresponding error image in the field due to the only error icon div getting hidden. In Microsoft Internet Explorer 11, the UI form contained the corresponding error image in the field, but only part of this image was visible due to changes of column/field width not being honored. These issues have been fixed by updating the CSS in the pzbase-grids.css file.
SR-D86200 · Issue 549032
Resolved tab appearance issues related to resolution support
Resolved in Pega Version 8.1.9
When trying to add multiple tabs on larger screens, once the first row was filled with tabs then new tabs began appearing at the bottom of the layout group. The layout group worked as expected with the tab overflow having the right and left arrows with the dropdown on smaller screens. In addition, @1440px the overflow arrow on the right side was not working correctly, showing as grayed out even though focus was not on the last tab. These issues were traced to taboverflow with add icon not working correctly at certain resolutions. To resolve this, the todisableslide function has been updated to consider add icon width and layoutgroup has been updated to consider parent width.
SR-B48818 · Issue 305799
Enhancement to make Cassandra data file location configurable
Resolved in Pega Version 7.3.1
The location of the Cassandra data directory is hardcoded in versions 7.2.1 and greater and points to a different directory than previous releases. This caused problems for upgrades from pre-7.2.1 systems where disk volumes may already be configured specifically to house Cassandra data in the old locations. For Tomcat installations, the data file location is hard coded to be a directory under the Catalina Home directory. In customer systems, this creates a problem because the home directory is frequently not configured to be big enough to accommodate the Cassandra data volume. In order to improve compatibility, a new configuration property (DNode/cassandra_data_directory) has been added to override the hardcoded location using prconfig.xml, Dynamic System Settings, or the command line. Configuration options are provided for prconfig.xml and Dynamic System Settings that allow specifying the location of the Cassandra data.
SR-B54923 · Issue 312483
TrimLog regression for BIX extract purge fixed
Resolved in Pega Version 7.3.1
After upgrade, it was noted that execution history BIX extracts were unexpectedly only kept for 30 days. SystemCleaner calls TrimLog for specific sub-classes of Log- (i.e., Log-UndefinedMessage, Log-Operator-Session, etc.) and a daysAgo parameter to specify the number of retention days to keep. Previously, TrimLog would select up to 200 of whatever Log- subclass that were older than the retention days and delete them. As an unintended consequence of performance improvements, a modification to TrimLog to call purgeTableContainingClass meant that while this activity does use daysAgo to delete only rows older than this many days, this method did not take subclass into account. This resulted in the API purgeTableContainingClass() truncating the entire table without adding pxObjClass filter. This regression in behavior has been resolved by modifying the TrimLog activity to do an executeRDB query instead of calling purgeTableContainingClass().
SR-B50813 · Issue 315070
Improved performance for Proposition updates
Resolved in Pega Version 7.3.1
While updating propositions through DSM portal it was observed that the thread count got very high while overall performance became degraded. This happened in Pega versions where the propositions are data instances and not rules. The issue was traced to the Proposition cache getting reloaded from the database table in a synchronized block; all of the strategy execution threads were waiting until the reloading completed despite strategy reading the propositions from the cache. To resolve this, the system will reload the proposition cache from the database table in the thread where the pulse message is being processed when the node receives the pulse notification. Thus other strategy execution threads will not be blocked and continue reading propositions from the cache.
INC-130695 · Issue 587659
Enhancements for upgrading in multi-tenant environment
Resolved in Pega Version 8.1.9
Some muti-tenant installations use the same applications or rule instances with the same pzInsKeys for different tenants. This can cause upgrades to time out due to the system fetching all pzInsKeys (which will have duplicates) and working with them in a default batch size of 500 each over 4 threads. This led to the same keys potentially being allocated and processed in different threads, resulting in duplicate processing and timeouts. This has been resolved by updating the select query to fetch the tentantid and pzInskeys in the MT system to avoid duplicate work in multiple threads. In addition, running Generate Declarative indexes fetches the pzinskeys and generates indexes for each record, but before generating, the existing index for the record is deleted and then inserted. Because the delete query to generate the index was not tenant aware, all of the records for the key were deleted for the tenants for that key, but the new index was created only in one tenant. This has been resolved by enhancing the DELETE query to be tenant aware, which will avoid deleting the indexes for all the tenants given an index key.