SR-B95350 · Issue 341577
Rollbacks use CreateDateTime for ordering list
Resolved in Pega Version 7.4
When processing a rollback for multiple hotfixes, the system was retrieving a list of uncommitted hotfixes from the database without specifying any order and, as a result sometimes the item were in the correct (i.e. reverse of installed) order and sometimes not. This was the unintended consequence of switching to using generated unique IDs which do not maintain any ordering, and has been fixed by modifying the process of retrieving the uncommitted hotfixes list to provide descending ordering based on pxCreateDateTime.
SR-C4008 · Issue 346439
Tile-generator updated to v1.8
Resolved in Pega Version 7.4
Tile-generator has been updated to version 1.8. This includes support for multiple subnets.
SR-C6074 · Issue 349037
Indexing doc clarifies JVM setting ramifications
Resolved in Pega Version 7.4
The Indexing section of the install / upgrade guides recommended setting the -Dindex.directory JVM argument, but did not explain the consequences of not doing so. The help document and PDN have now been updated to clarify that if this setting is not used and if a custom index file directory is used on the Search Settings Landing page, the system will revert to using the default directory if no other index host nodes are online when a server starts.
SR-C6941 · Issue 343726
Rollback information collection optimization
Resolved in Pega Version 7.4
During rollback information collection, a rule inventory is regenerated. This checking can include searching for instances that no longer exist in the database, potentially causing performance issues. In order to optimize the function and reduce unneeded logging, the rollback information collection process has been updated to reuse the rule inventory.
SR-C8467 · Issue 349659
Support added for Cassandra on SunOS
Resolved in Pega Version 7.4
Changes have been made to startup scripts and lz jar version in order to support C* on SunOS.
SR-C8702 · Issue 349662
Fixed cataloging for framework hotfixes with non-java code
Resolved in Pega Version 7.4
The system was unable to catalog Strategic Application hotfixes that included non-java code instances (e.g. .xml or .properties files) due to Framework hotfixes using a different version format that caused validation failure. To correct this, the system will bypass redundant release version validation for Strategic Application hotfixes so that it can catalog non-java code instances.
SR-C8702 · Issue 344809
Fixed cataloging for framework hotfixes with non-java code
Resolved in Pega Version 7.4
The system was unable to catalog Strategic Application hotfixes that included non-java code instances (e.g. .xml or .properties files) due to Framework hotfixes using a different version format that caused validation failure. To correct this, the system will bypass redundant release version validation for Strategic Application hotfixes so that it can catalog non-java code instances.
SR-B69324 · Issue 327826
Support added for removal of empty ComplexTypes entries
Resolved in Pega Version 7.4
When using several complextypes in Stream XMLs for Connect REST integration, empty values were being passed automatically for ComplexTypes that are optional. Support for the removal of empty entries from the PageList/PageGroup property has now been added.
SR-B73008 · Issue 329204
Support added for two-letter tenant names for REST
Resolved in Pega Version 7.4
Due to HTTP/REST services using the last two characters of a token to identify whether the token was a tenantHash or a service Package, REST did not work when accessed by a tenant with a name consisting of only two characters. As comparison with only the last character is required, the extractKeys API of HTTP Service has been updated to use just the last character of a token to identify if the token is a tenantHash or a service Package.
SR-B73208 · Issue 329189
Support added for two-letter tenant names for REST
Resolved in Pega Version 7.4
Due to HTTP/REST services using the last two characters of a token to identify whether the token was a tenantHash or a service Package, REST did not work when accessed by a tenant with a name consisting of only two characters. As comparison with only the last character is required, the extractKeys API of HTTP Service has been updated to use just the last character of a token to identify if the token is a tenantHash or a service Package.