SR-D5076 · Issue 456380
Table layout overflow fixed for IE11
Resolved in Pega Version 8.2.3
An update made to address an IE11 display issue when a table layout had more than 10 column by adding a display:inline-block caused unforseen issues. The inline block has now been removed and a permanent solution to the display issue will be available in a future release.
SR-D8390 · Issue 488001
Handling added for runtime exception on upgrade due to inaccessible external schema mapping
Resolved in Pega Version 8.2.3
If a system used external schema mapping, attempts to upgrade were failing with the error "Caused by: com.pega.pegarules.pub.database.ConfigurationException: Unable to obtain data source "java:comp/env/jdbc/PegaRULES", as it is undefined". This was traced to configuring external DADNs which were not accessible from the server where the upgrade was being run: an upgrade will fail if it cannot connect to an external datasource even if no changes need to be made to that datasource. However, because a PRRuntimeException was unexpectedly being thrown instead of a Database Exception, the exception was not properly caught and handled. This has been corrected.
SR-D6978 · Issue 481778
Mandatory field confirm popup text and focus corrected
Resolved in Pega Version 8.2.3
When a mandatory field is tabbed out (with TAB key), using cancel to proceed to the next field instead retained the focus in the same field. This was traced to work previously done to prevent an infinite confirm dialog loop, and has been corrected.
SR-D24261 · Issue 492557
Updated retry for context registraton
Resolved in Pega Version 8.2.3
The system was intermittently hanging after importing zipped files. Investigation showed that Batch threads were becoming stuck in AuthorizationContextManagerImpl.setSpecializations() due to a 'while' loop in setSpecializations that was seeking to register the new context. To resolve this, logic has been added which will try to register up to 10 times. If for any reason it can't register, the system will just return the unregister LAC. The check before deregister has also been enhanced and now will only deregister if the new context and the current context are different. It will not call register if the current context and the new context are the same.
SR-C80200 · Issue 494447
Resolved Sporadic ControlBuild error in sections
Resolved in Pega Version 8.2.3
After upgrade, a sporadic error was seen with sections of a VTable after a server restart where a portion would not load and the error "com.pegarules.generated.html_property.ra_stream_pxtextinput_1786000787_default incompatible with com.pega.pegarules.pub.runtime.ControlBuilder" was seen. This issue was traced to invalid generation of an assembly parameter map for random sections the first time the table was called after restart. To resolve this, the system will now allow assembly parameter generation for a second time when they are empty, and the assembly parameters preparation will be handled as a synchronous process.
SR-D28905 · Issue 500673
Compensation added for Cassandra v2.2 not honoring count limits
Resolved in Pega Version 8.2.3
Running CassandraEstimateNumberOfRecordsOperation on a large data set was causing a ReadTimeoutException. Investigation showed that contrary to its documentation, Cassandra v2.2 ignores limits intended to stop all records from being counted. In order to compensate for this, the system has been modified to skip the first of two attempts to estimate the number of records (first cql count() query with limit 1000) and the estimated number of records will only be retrieved through JMX.
SR-C96680 · Issue 483765
Reject changes in approvalRuleChanges flow will clear cache and reassemble rules
Resolved in Pega Version 8.2.3
When a paragraph rule with "need approval" checked in the ruleset version was delegated, previewing the data in the paragraph rule via a button and making changes worked as expected. However, after the changes were made and sent to the approver to accept or reject the changes and the private copy was deleted, it was observed that rejected changes still showed in the paragraph rule and the cache was not cleared. The Rule Approval process allows checkouts to be approved or rejected: when not using rule approval process, a Checkin from the personal ruleset results in a Delete operation (from WBCheckIn & RuleCheckIn) that will remove the rule from the personal ruleset. However, with Rule Approval process setup, the rule in the personal ruleset is not deleted. Instead it is simply moved to the CheckInCandidates ruleset. The pyRuleset on the rule page changes, but the Inskey remains the same. Because there was no delete operation, there was no noteChange event and the cache was not cleared. To resolve this, the system has been updated to now enable NoteChangeInternal to invalidate rules that are part of the personal ruleset only if they have the correct setting in their prconfig.xml. This fix also forces re-assembly of rules after check-in and approve/reject so that pressing preview will ensure the correct rule gets displayed.
SR-D16829 · Issue 486655
Nodes start as new if replacement start fails
Resolved in Pega Version 8.2.3
Cloud upgrade was failing if there was any attempt to scale up the number of instances. Every node starts up with a flag, and in cases of scale-up scenarios, a new instance with this flag was handled as a replacement instance instead of a new instance. That led the new node to fail because it has nothing to replace. To resolve this, the system has been updated so that if an instance failed to start as a replacement, the process will allow the instance to join the cluster as a new node.
SR-D17741 · Issue 486919
Nodes start as new if replacement start fails
Resolved in Pega Version 8.2.3
Cloud upgrade was failing if there was any attempt to scale up the number of instances. Every node starts up with a flag, and in cases of scale-up scenarios, a new instance with this flag was handled as a replacement instance instead of a new instance. That led the new node to fail because it has nothing to replace. To resolve this, the system has been updated so that if an instance failed to start as a replacement, the process will allow the instance to join the cluster as a new node.
SR-D17743 · Issue 486915
Nodes start as new if replacement start fails
Resolved in Pega Version 8.2.3
Cloud upgrade was failing if there was any attempt to scale up the number of instances. Every node starts up with a flag, and in cases of scale-up scenarios, a new instance with this flag was handled as a replacement instance instead of a new instance. That led the new node to fail because it has nothing to replace. To resolve this, the system has been updated so that if an instance failed to start as a replacement, the process will allow the instance to join the cluster as a new node.