Troubleshooting flow execution

To troubleshoot flow execution problems, follow these steps:
  1. Run flow with audit, debug and trace on using flow command line options: -ad --trace INFO
  2. Examine the flow audit file.

If the flow fails at the node level, look for the first_failure json key to get the information about the node that failed. For example, if a mapping node fails to get the input file, an example first_failure key value is:

"first_failure":{"node":"MyFlowName::MyNodeName","message":"Source not available","code":-48}

To get details about the failure:

To get details about the failure, find MyNodeName in the array of nodes in the audit file. The detailed message for the map audit will be the following:

"Message": "Source not available",
"CommandLine": "c:\\deployment\\maps\\MyMapName.mmc",
"SourceReport": [ 
        {
        "card": "1",
        "adapter": "File",
        "bytes": "0",
        "adapterreturn": "-1",
        "Message": "Source not available",
        "Settings": "c:\\deployment\\input\\input.txt"
       }
…

Examine the flow trace file from the logs folder:

A message similar to this one will be reported in the flow trace file:

[ 10/01/20 16:27:52 ] UUID: a57384e3-7448-494c-91ca-6a478497efeb RC: 0 ID: 4136 Message: DTXLN4136E tid= 26976 inst: 1 - I/O ERROR : I/O type=0 Open Map= c:\deployment\maps\MyMapName.mmc Failed to open c:\ deployment\input\input.txt Error=2 (mercio.c:2234) Flow Name: MyFlowName::MyNodeName

If a flow fails with a catastrophic error, such as Segmentation Fault on Windows, or aborts with SIGSEGV, SIGABORT, SIGBUS or SIGILL on Unix, a crash log file is created in the logs folder. The crash log file name format is:

flowcmdserver_crash_PID_datetime.log.PID

HCL Link support might ask for the crash log to help diagnose the problem.