Fixed Various Issues on CDC Steaming API #
- There was an issue when the file was opened and the Salesforce CDC raised an event. A new message was created, but the record was not inserted. This issue has been resolved by enabling data to be written to the file during the file opening and updating the message status to ‘completed’ after finishing the record insertion.
- The previous version’s event message did not provide data for monitoring. We have now fixed this by adding data to the event message.
- When you update a field on a record that is not mapped, many messages are created. This issue has been fixed by no longer creating and sending these empty records to the file.
Fixed Store Procedure Postgres Does Not Support the Parameter #
The problem is Postgres doesn’t support the parameter of the store procedure, so we need to set it as a column index. The most important thing is we need to tell the connection that we will store procedure NOT function like this “escapeSyntaxCallMode=call”. Example: “jdbc:postgresql://localhost:5432/postgres?currentSchema=sobject&escapeSyntaxCallMode=call”.
Fixed Error while using Reference Field on Agent File Adapter with SFTP Protocol #
We encountered an error when using the Reference Field on the Agent File Adapter with the SFTP protocol. The error log indicates that ‘the specified system path could not be found. Now this issue has been resolved by working on the Reference Field.
Message Status Not Updated After Exporting the Binary File #
This error occurred while you were exporting the binary file. You received the file, but the message status remained ‘New Status.’ We have now fixed the issue, and the status will be marked as ‘Completed’ once the file is exported.
Fixed Missing Agent Password on Cache #
There was an issue with the password cache missing in the salesforce_authentication table in Postgres. We have now fixed it by installing the required field to prevent any missing values.
Fixed Response Scheduler Error to Scheduler Screen on Agent Control Board #
In the previous version, we did not properly identify scheduler errors. We have fixed this by providing an error message to report the error from the agent side.
Fixed Scheduler When Service is Stopped So Making The Next Run Invalid #
We encountered an issue where the scheduler’s current DateTime was less than the Next Scheduled Run, causing the Trigger States to be in error. This issue has been fixed to prevent it from happening again.
Logging not Able to Understand while No Data to Do Integrate #
This issue occurred when the query did not return any records to be integrated, as there were no records in the database that matched the query criteria. The log message ‘dataIbean is loaded. Size of dataIBean: 0’ was unclear. We have now resolved this by logging the message as ‘No Data Available in [Interface Name]’.”
Bulk Basket Did not Create With V2 Interface Design #
The bulk basket is not created when using Skyvva Bulk with Interface V2 Interface Design because the New Agent creates an attachment instead of a content version. This issue has been resolved by creating the bulk basket with the ContentVersion.
Fixed Cache Issue #
Fixed some issues with synchronizing between Salesforce and the agent cache.
Fixed Inserting Hierarchical Records into Postgres Database #
We encountered a problem while inserting hierarchical records with a parent root map where the primary key is an auto-increment number. The issue is that we did not exclude the primary key from being mapped to the database. For example, the id field is the primary key in the database and is mapped to the Salesforce ID in Salesforce. This field is intended to be used as a foreign key in the child table only. However, we don’t need to insert it into the database because it is auto-incremented. Therefore, when the child table is inserted into the database, it will reference this ID field as a foreign key. Now this issue has been resolved.
Fixed Issue Not Able to Update the Message on Agent File Adapter #
- There was an issue with outbound processing using the Agent File. You received the file, but the message status did not update and remained as ‘new’. The problem was that the local file adapter did not cache messages that were requested to be updated. This issue has been resolved by updating the status to ‘completed’ once the Agent receives the file.
- When exporting the file attachment, we initially received a new message and did not get any file attachments. We corrected it and reprocessed the messages, but the message status remained ‘new’ and did not update. However, the file was received. This issue has been fixed by correctly updating the message status.
Fixed Support sObject Message Type with Agent File Adapter #
In the previous version, the new agent did not support the V3 sObject Structure. Now, we have fixed it to support the V3 sObject Structure for both inbound and outbound Agent File Adapters, making it easier for users as they have the same structure as Salesforce.
Issue: Running Parallels Manual and Scheduler Simultaneously #
There is an issue when performing manual processing while the scheduler is also running to execute the same interface. This results in the interface being executed twice simultaneously. The processing type does not update correctly and remains Manual Processing. This issue has been fixed by checking if the first sequence of the interface is done by manual processing, and the scheduler will then check the second sequence of the interface to be executed.
Fixed Issue Backup File While Processing Failed #
We encountered an issue with an incorrect CSV file that we used to import via the Agent File adapter. The file could not be imported due to an error message: ‘Exception Caught {} fboundsException: Index 4 out of bounds for length 4,’ which was caused by missing commas in the file. Despite this, the file was still backed up. This issue has been resolved by ensuring that files are not backed up when file processing fails, and by improving the error logging for better understanding.
Inbound V3 With Agent File Adapter Support a Source to Many Target Mapping #
In the previous version, the inbound V3 Agent File adapter did not support a source node map with multiple target nodes, causing the child target node map to miss inserting child records. This issue has been resolved by supporting multiple target nodes to insert hierarchical records.
Fixed Consumer File Inside the Adapter #
In the previous version, we stopped the consumer on the adapter, but the file was still being imported. This issue has been resolved by changing the status check from ‘Status’ to ‘Consumer Status’ to verify the consumer’s status.
Fixed Getting Two Responses to Update While Having a Request Insert Record into Database #
There was an issue when sending a request to insert a record into the database via the Agent Database Adapter, resulting in two responses to update the record in Salesforce. This issue has now been fixed, and you will receive only one response per request.
Fixed ODBC with Datatime Format #
In ODBC, we encountered an issue where inserting a null value into a DateTime field column in the database resulted in the value ‘1900-01-01 00:00:00.000’ being inserted instead. This issue has now been fixed, and inserting a null value will no longer result in this default value.
Stored Procedure Issue Regarding Requests Without Responses #
Fixed the outbound interface with the Agent Database adapter using a stored procedure. Now, the request interface no longer requires a response interface, making it optional. If there is no response interface, we will process only the IN parameter. Additionally, we have switched from using parameter names to using indexes instead.
Added All Tables to View Agent Cache while needing to View #
We cannot check some agent caches on the Agent Control Board. For example, we cannot see the qrtz_triggers table to determine why the scheduler is not running. We have now added all agent cache tables to ensure we can view all necessary tables.
Fixed Issue Allowing Interface That is Development Status #
In this case, although the interface was not deployed, the agent still executed it, which led to unexpected behavior. This issue has been fixed by not allowing the execution of interfaces that are still in development status and by verifying the deploy status of interfaces before execution to avoid such errors.
Fixed Different Job Status Between Manual Processing And Scheduler Processing of New Agent #
There was an issue where the same interface was executed via New Agent, but the Job Status differed between Manual Processing and Scheduler Processing. We have now fixed this issue so that the Job Status is consistent in both cases.
Fixed Different Job Status Between Manual Processing And Scheduler Processing of New Agent #
When you set the New Agent Inbound Scheduler, the Outbound Interfaces were executed. This issue has been resolved by preventing the New Agent Scheduler from executing Outbound Interfaces.
Fixed Include timestamp into file name for SFTP and FTP File Using backup File #
In the previous version of the Inbound Agent File Adapter, when you backed up the file after processing, it did not include the ‘processed’ tag and timestamp. We have now fixed this by concatenating the original file name with ‘processed’ and a timestamp (ZonedDateTime.now(ZoneId.of(‘UTC’)).normalized().toString()).
