SR-129260 · Issue 196814
Improved passivation handling for long page names
Resolved in Pega Version 7.1.9
An issue where sporadic 'database serialization failed for requestor' errors were being logged was traced to certain page names that are very large. During passivation, if the pxInsName was larger than 128 characters, it did not fit in its associated column in the pr_page_store table and generated the error. It was also noted that while pzinskey is 255, pxinsname is 128. To resolve this error, pxinsname has been increased to 255, and DatabasePassivation will hash the page name for storage if it exceeds the maximum allowable length.
SR-129984 · Issue 203634
Error handling added for skin check-in
Resolved in Pega Version 7.1.9
A 'string index out of range' error when checking in a skin was causing the rule to hang. This was caused by an exception generated by the ordering of clipboard results, and handling for this has been added to resolve this problem.
SR-130496 · Issue 203385
Fixed ruleset cache population after passivation
Resolved in Pega Version 7.1.9
After a session was idled, a 'RulesetName not specified' error appeared during checkout when the session was resumed. This was an issue with SnapshotApplicationContext not being activated properly to populate the cache after passivation, and has been fixed.
SR-130754 · Issue 200600
Remove locale consideration from toDecimal string conversion
Resolved in Pega Version 7.1.9
The 'toDecimal' function was unable to convert the string to decimal properly when the user locale is set to Indonesia. As the 'toDecimal' RUF should not be doing localized conversions, this RUF has been updated to ignore user locale.
SR-131553 · Issue 203114
Passivation handling added to changelist
Resolved in Pega Version 7.1.9
A Rule-Declare-Index defined on a Context page property was intermittently failing to fire. This was traced to a missing changelist following passivation. Changelists were never originally designed to be passivated, but it is possible to passivate requestor or PRThreads which then affected the pages with changelists. The system has now been updated to properly passivate changelists in order to prevent this issue.
SR-131691 · Issue 204727
Improved performance for incorrect password handling
Resolved in Pega Version 7.1.9
When the operator entered the wrong password in the login screen, numerous "Stream Overwritten" alerts appeared the Alert log. There was no workflow problem involved, but the logging indicated that there was an additional unnecessary call being made to Stream Web-Login, and that unneeded call has been removed to improve system performance and remove the error being logged.
SR-131726 · Issue 202769
Fail-over updated to cover server crash scenarios
Resolved in Pega Version 7.1.9
In a testing event, the user session did not properly fail over to another server if the original server was shut down; an error was thrown instead. This was found to be an authentication issue where on node crash, a new requestor was created on the second node and then immediately authenticated as it is SSO. As part of authentication, the mRequiresSpecialProcessingForNodeCrash flag was cleared, meaning the node crash activity was not executed and entire portal was not refreshed. This has been corrected by updating the system to execute crash node activity for SSO cases.
SR-132732 · Issue 201617
SMA Instance MRU report updated
Resolved in Pega Version 7.1.9
The SMA Advanced / Rule Cache Summary report showed "NA" for pruning, even when the system had been impaired by extensive pruning. The correct data was available under Advanced / Reports / Most Recently Used Caches Report, but the NA results were purposeful: this particular report branch had been incompletely refactored while implementing multitenancy, and the decision had been made to have this report show NA rather than zero because zero was misleading and inaccurate. We apologize for any confusion this may have caused, and the report should now accurately report the expected results.
SR-132809 · Issue 207385
Next run in agent schedule updates after DB access issue
Resolved in Pega Version 7.1.9
It was discovered that if access to the database was interrupted, the Next run in agent schedule was not updating and not being run. This has been fixed by updating AgentQueue.handleProblem() to handle unexpected errors and ensure delayDueToProblem() is invoked always.
SR-133057 · Issue 205137
System updated to better handle multi-node rule resolution
Resolved in Pega Version 7.1.9
In an upgraded multi-node environment, the error 'com.pega.pegarules.pub.database.MultipleRuleVersionException: Rule resolution identified 2 versions of the rule' was sporadically appearing. Once the issue occurred, it would persist and the only way to resolve it was to bounce the server. This was an issue with an instance of one rule being found in one node and not found in other node, and was related to the trigger logic. This should not occur with the current Pega7 version, as trigger logic has been moved into the engine and there is no longer a database trigger to fix. In order to ensure continued smooth trigger resolution, tests have been added that will check for this condition in the updates cache table to ensure that this scenario does not occur again