Long running publish job abruptly errors out during remote publish.
Each executor in the remote engine has a timeout setting. Examples of executors are GTDatamaker.exe, Testmatch.exe, ALMBatch.exe etc. These time outs come into the picture if a remote job of any of the executors run beyond a stipulated period of time after producing .ERROR file. Unlike other executors, GTDATAMAKER.exe produces an .ERROR file at the start of the publish process and replaces it with an .OK file on successful completion. This is the reason remote engine starts counting GTDatamaker's execution time right from the start. By default, the timeout setting for GTDATAMAKER's executor is 7200 seconds, which is 2 hours. This is usually sufficient for most jobs except when GTDatamaker is used to generate more than 80,000 rows such as that of a Test Mart. In these cases, the execution time of remote publish tends to exceed 2 hours and the remote engine erroneously kills it.
Test Data Manager - Remote publish mode
If such a long running publish process is used to populate a Test Mart, then the recommendation is to use Javelin instead. If Javelin cannot be used or if the remote publish is actually generating test data, then it is advised that the timeout setting of GTDATAMAKER executor is increased suitably from its default value of 7200 seconds.
1. Navigate to the installation folder of Remote engine usually found at C:\Grid-Tools\RemotePublish.
2. Open the DMBatch.exe.config file in an editor.
3. Change ShutDownTimeIfJobHangsOnError from 7200 to a suitable time (for example 20000) against the following 3 entries.
4. Restart Remote engine. (Either from services.msc or the config editor)
If you face any issues or have any questions about the steps, please open a support case by going to https://support.ca.com/irj/portal/anonymous/newhome or by calling customer care at http://www3.ca.com/us/support/customer-assistance.aspx.