Release notes for RunMyJobs releases 2025.1.0.0 through to 2025.1.0.7

Before: Invalid HTTP REST calls into RMJ may throw FATAL log messages

After: Invalid HTTP REST calls into RMJ will not result in FATAL log messages and are handled gracefully.

Before: when a user chooses to execute the action “Run Now” on a Process that is scheduled to run in the future, the system adds a note to that Job. The message in this note cannot be longer than 1000 characters. Under certain circumstances the message could grow beyond this limit causing an error.

After: the message will be limited to 1000 characters. In case it is longer, the last 3 of those 1000 characters will be replaced by dots (a.k.a. ellipsis) to indicate that there was more text that didn’t fit.

Before: System_RaiseClearEvent Job supports only Raise and Clear actions.

After: It is possible to ‘Clear all pending’ and ‘Clear Including pending’ raised events using System_RaiseClearEvent Job.

Before: Custom Libraries incorrectly used stale compile jar files, when they included REL Entrypoints, and all the supplied sources were replaced by uploaded jarfiles.

After: Custom Libraries correctly remove stale compile jar files.

Before: Reading excel files with empty lines was incorrect.

After: Reading excel files with empty lines was improved.

Before: The linked file fails to open after the job completes and throws an exception.
After: The linked file should open successfully without throwing any exception.

Before: Creating a proxy credential with a missing or null hostname resulted in no UI error and a startup failure with an error message in the scheduler.log file.
After: Creating a proxy credential with a missing or null hostname now displays the error message "The proxy URI host is missing" on the UI, preventing startup failures.

Before:  If you want to submit jobs in a negative timezone (E.g. America/New-York) using the Dates option (Recurrences) in the future month the jobs are calculated in the wrong month and you can’t save the recurrence.

After: This has been fixed.

Before: SQL exception in report job definition due to obsolete CategoryProcessTag field in legacy imported report

After: SQL exception resolved in migration if legacy report includes CategoryProcessTag column

Before: RMJ was running on Tomcat 9.0.87, which is vulnerable to CVE-2024-38286, CVE-2024-34750, CVE-2024-52316, CVE-2024-52317 and CVE-2024-52318

After: RMJ is now hosted on Tomcat 9.0.97, which is the latest release and without known vulnerabilities.

NOTE: The configuration of RunMyJobs limits the impact of these vulnerabilities, making the believed risk of exploitation low, however the version is being upgraded out of precaution.

Before: RMJ was running on Tomcat 9.0.97, which already includes a fix for CVE-2024-52316 but remains vulnerable to CVE-2024-50379 and CVE-2024-56337.

After: RMJ has been upgraded to Tomcat 9.0.98, the latest release, which addresses these vulnerabilities, potentially mitigating the associated risks.

Note: While the configuration of RunMyJobs significantly limits the potential impact of these vulnerabilities, making the likelihood of exploitation low, the upgrade to the latest version of Tomcat has been carried out as a precautionary measure to enhance security.

Before: In some cases when upgrading a Library, and there are a large number of Jobs in the system that use that Library, a large amount of memory would be used, which could even cause the update to fail.

After: The amount of memory used during the update of a library is dramatically reduced.

Before: Executing System_Mail_Send job with an attachment produced by a platform agent resulted in creating a new JVM thread that would stay in waiting state until the server was restarted. Beside causing thread pool exhaustion, this could also lead to OutOfMemoryError.

After: When reading mail attachments the input stream is properly closed and the JVM thread is returned to the pool.

Before: Variant description was not imported with the variant

After: Variant description is imported with the variant using the language of the RFC user configured for the client

Before: Platform Agent and jtools still reported old ciphers and older security protocols sslv3, tls10 and tls11.
After: Platform Agent and jtools do not report (nor support) old and obsolete ciphers and security protocols (sslv3, tls10 and tls11) anymore.

Before: SQLPLus jobs only determines tnsnames.ora from ORACLE_HOME/network/admin or SQLPLUS DIR/network/admin.

After: From Oracle 18 and up Oracle uses a different path the set the tnsnames.ora. It uses ORACLE_BASE_HOME/network/admin to set the tnsnames.ora file. We should also detrmine if we have an ORACLE_BASE_HOME.

Before: Platform Agent on some VMS could potentially fail at job completion.

After: Platform Agent on VMS will no longer potentially fail during job completion.

Before: We are sending a SOAP request to SAP Data Services even if the session is null. This is the correct scenario only for a Logon request.

After: If the session is null and it is not a Logon request, an exception will be thrown, and no SOAP request will be sent.

Before: Opening the Studio Job Library could initially take a long time.

After: The Studio Job Library now uses “infinite scrollers” for both Folders and Jobs which take considerably less time to load and fetch more when scrolled to the end.

Before: It was not possible to connect with SSL to AS400.

After: When setting ProcessServer Parameter SecureConnection to true we will connect with SSL to AS400.

Before: There's no way to configure a workflow to wait for a job within it, except to wait for its step.

After: The 'Workflow waiting on job' flag is introduced at workflow job level, when set to true the workflow will wait for the job to complete. This value will be ignored if 'Step waiting on job' is set to true.

Before: When a chain with 2 steps was run at 10PM, job at step 2 has a scheduling parameter start time set to 11PM and jobs in step 1 took longer then an hour, the job at step 2 got rescheduled for another day.

After: Introduced UseRelativeTime flag on job chain call, when set to true, the start time of the chain is used to evaluate job scheduling parameters (time window, start time) and preconditions (time window).

Before: When the HTTP call fails and the External Completion Strategy was set to external the job stays in Running state because the jobs wanted to be updated by an external system. However the external system could not be aware of the job running.

After: When the HTTP job gets a HTTP status code between 200 and 300 the External Completion Strategy is set. Outside these status codes the job will be put to Status Error and if http messages exists this will be printed in the stdlog.log file

Before: There was a limit of 255 characters for the Job alert source, Process server alert source, and Queue alert source.

After: Alerts Name Pattern field supports up to 1024 characters.

Before: System_RaiseRemoteEvent doesn’t support Remote System as a connection parameter.

After: New System_RaiseRemoteSystemEvent Job supports Remote System as a connection parameter, System_RaiseRemoteEvent is deprecated.

Before: If duplicate objects (objects with the same business key) are detected during the import, they are ignored.

After: A new configuration is introduced: ImportingConfig DuplicateObjectHandlingOption, if set to IgnoreDuplicates the system will behave as before, if set to ErrorOnDuplicates the import will fail with an error if duplicates are detected.

Before: Communication between RMJ and the PlatformAgent (Server Initiated) was managed through a custom HTTP communication implementation.

After: To improve security, we updated the communication to use Java’s standard HttpClient. Additionally, for SSL connections, we now rely on Java's built-in certificate validation mechanism. ProcessServer parameter CheckAgentCertificate is now deprecated and has been replaced by the registry setting "VerifyServerCertificateTrust" with the default value set to “Verify”.

If you use a secure connection between RMJ and the PlatformAgent and the certificate had not been imported yet, the ProcessServer may fail to start. In such cases, set the "VerifyServerCertificateTrust" registry parameter to Warn. This will check the certificate but allow the ProcessServer to start and an operator message will be generated if the certificate cannot be verified.

Before: Out Of Memory issues could occur when transferring files between a Platform Agent and a Central Server

After: Optimized copying of files while transferring files between a Platform Agent and a Central Server reduces chances of Out Of Memory issues

After:

Allows to download the PSE Certificate Request (sapsnc.csr).

Allows to upload the certificate response (sapsnc.cer).