SR-A95518 · Issue 264095
Resolved loop in batch insertion
Resolved in Pega Version 7.2.2
If due to any reason (ex, duplicate fact id or database crash) a batch insertion is not successful, the system should retry 5 times before an error is thrown. An issue with the retry logic was found that caused a flag once initialized to 'true' to not ever be set to 'false', leading to a loop situation that inserted duplicate data. This has been corrected by setting the retry flag to false once insertion is successful (failed=false) so that loop exits.
SR-A87133 · Issue 255127
Allowed import of RAP containing ADM
Resolved in Pega Version 7.2.2
Import of a Pega Marketing .rap file containing Adaptive models failed if an ADM host was not available during import. This was caused by the system checking if the server had that Rule (configuration) already before processing the save request, and has been resolved by adding an extra test in pxOnDoSave activity to not check for existing configurations if there is no ADM node configured.
SR-A87739 · Issue 257135
BeanShell upgraded for security
Resolved in Pega Version 7.2.2
In order to address a potential security vulnerability in BeanShell that could be exploited for remote code execution in applications that have BeanShell on its classpath (CVE-2016-2510, BeanShell has been upgraded to org.apache-extras.beanshell:bsh:2.0b6
SR-A102021 · Issue 268416
Cassandra keyspace configuration updated
Resolved in Pega Version 7.2.2
In order to better support external Cassandra instances and using DDS-based Cassandra when the logic is running from a non-DDS node, the datacenter name will be read from Cassandra using the Datastax driver rather than JMX. The Datastax driver will request the information across the network if needed, rather than trying to use the JMX connection to Cassandra on the localhost (which would fail since there is no JMX port open because there is no running Cassandra on the non-DDS Node). In addition, creation of the keyspaces will be performed on the startup of the first DDS instance prior to any other DSM services running. This will prevent the other services from attempting to create the data keyspace since it will already exist.
SR-A93286 · Issue 270071
Data flow shape batch size increased from 1 to 250
Resolved in Pega Version 7.2.2
In order to support large and complex systems, the data flow shape batch size has been increased from 1 to 250.
SR-A93286 · Issue 265355
Data flow shape batch size increased from 1 to 250
Resolved in Pega Version 7.2.2
In order to support large and complex systems, the data flow shape batch size has been increased from 1 to 250.
SR-A76054 · Issue 251761
Enhancement added to bulk delete ADM models
Resolved in Pega Version 7.2.2
In order to support the deletion of ADM models in bulk, an enhancement has been added to remove the unwanted models from the database provided a list of them is given as input. The new activity is pxDeleteModelsByCriteria, applying to DSMPublicAPI-ADM. The pzInsKey of this activity is: RULE-OBJ-ACTIVITY DSMPUBLICAPI-ADM PXDELETEMODELSBYCRITERIA #20160608T102848.043 GMT It has been created in ruleset Pega-DecisionArchitect:07-10-16 on 7.1.7 HFIX system vengwindb180:8282/. The usage text, explaining the activity, is as follows: "Used to delete ADM models in bulk. Models to be deleted are determined by the criteria selected by this Activity's parameters; a model is deleted if it matches all selected criteria. PLEASE NOTE: activity is not constrained by Application, only by the criteria provided as parameters. Therefore in most cases you will probably want to constrain by 'applies to' class in addition to the other criteria. Please see the tooltip / description of each parameter for more info on their usage. Integer output parameter 'NumberDeleted' returns the number of models that were successfully deleted."
SR-A92631 · Issue 270068
VBD Actuals Query works correctly on WebSphere 8.5
Resolved in Pega Version 7.2.2
Nested JDBC metadata calls were not working in WebSphere 8.5/DB2 due to the nested metadata query causing the initial query's ResultSet to close prematurely. This was traced to unexpected behavior in IBM WebSphere 8.5 when querying DB column metadata while iterating another DB metadata ResultSet, and has been resolved by modifying the DSM/VBD code that queries the DB for dimension metadata so it will load the metadata in two stages to avoid nesting of metadata queries.
SR-A92952 · Issue 262158
Proposition filter validates all rows on save
Resolved in Pega Version 7.2.2
On changing the name of the strategy component , the proposition filter was not validating the change for all the propositions but was only validating for the first proposition which had focus. The initial design was intended to skip validation when the proposition row was not viewed by the user with the goal of saving time by skipping the propositions that were not edited. However, this posed a problem that if the referenced rules were removed then the rule continued to pass validation. To resolve this, the system will force validate all the rows even though they may not have been edited.
SR-B4681 · Issue 274245
Twitter quoted status generation updated
Resolved in Pega Version 7.2.2
Previously, generating connector metadata for Twitter retweet status used a sample of 100 random tweets. However, those 100 tweets may not have any retweets among them. In order to guarantee a retweet will be captured for connector metadata, the approach has been updated to look at the last 100 user tweets for generating the metadata; if any tweet is quoted in the last 100 tweets, quoted status will come in. This can also be used for other properties like image etc.