Skip to main content
Question

RTS 114 (Signal 11) in ESORAXA during PLTPI Phase 1 – Enterprise Developer 11 (32-bit) & Oracle 19c

  • June 9, 2026
  • 2 replies
  • 25 views

sergi.yague
Forum|alt.badge.img+1

Environment:

  • Product: Rocket Enterprise Developer 11.0.0.84
  • Platform: Windows 11
  • Architecture: 32-bit (Region configured with mfCAS64Bit: 0)
  • Database: Oracle 19c (Instant Client NT 19.30, 32-bit)
  • Region Type: Batch (JCL), although CICS personality is currently active.

Problem Description: I am experiencing a critical memory failure (RTS 114 / Signal 11) in the XA switch module ESORAXA during the startup of my region PROVARE5. The region initially reports that the XA interface is loaded and initialized, but it crashes immediately when reaching the PLTPI Phase 1 (XA recovery).

This crash leads to a CICS PIP1 transaction abend (ASRA), forcing the entire region to shut down before it can finish initialization.

Current Console Log Snippet:

CASXO0015I DBP XA interface initialized successfully
CASSI4104I PLTPI Phase 1 starting XA recovery program
CASSP0014S Process termination due to RTS error. Execution error: file 'ESORAXA'
error code: 114, pc=0, call=1, seg=0
114 Attempt to access item beyond bounds of memory (Signal 11)
CASKC0041E Transaction PIP1 abend ASRA. The task has terminated abnormally because of a program check.
CASKC0052S One or more core process(es) terminated prior to initialization - shutting region down

Steps Taken So Far:

  1. Switch Module Compilation: Compiled ESORAXA.cbl using the build ora19 utility from the Enterprise Developer Command Prompt (32-bit). The build was successful, and CBLLINK correctly used the Microsoft Incremental Linker [Log del usuario].
  2. Open String Refinement:
    • Reordered the string to place Oracle_XA at the beginning.
    • Current string: Oracle_XA+Acc=P/user/password+SqlNet=DBD_LINUX+LogDir=C:\Oracle\XALogs\PROVARE5\esxalog+SesTm=100.
  3. Clean Start: Attempted multiple "Cold Starts" (casstart /rPROVARE5 /c) to clear shared memory and control tables [Historia previa].
  4. Workarea Cleanup: Checked for physical esxalog* trace files to delete them, but found none, as the module crashes before creating the first log entry [Conversación previa].
  5. Library Verification: Verified that the Oracle Instant Client (Basic, SDK, Precomp) is 32-bit and its /bin and /sdk/lib paths are correctly set in the system PATH and LIB variables [Log del usuario].

Specific Questions for the Community:

  1. Hardcoded DLLs? Examining ESORAXA.cbl, I see conditional references ($IF XAPLATFORM = "WINDOWS") to ORACLIENT19.DLL and ESXAEXTCFG.dll. Since Oracle 19c uses oci.dll, is it possible that ESORAXA is failing while trying to dynamically load a library name that doesn't exist?
  2. Batch vs CICS: This is a Batch-only region. Why is the CICS PIP1 transaction and PLTPI recovery phase interfering with the XA initialization if I only intend to run JCL? Should I use the BATCHONLY parameter or is there a recommended way to bypass CICS recovery for Oracle XA in JES regions?
  3. RTS 114 at PC=0: The error report shows pc=0. Does this typically indicate a null pointer during the xa_open or xa_recover call within the Rocket-provided switch source?

2 replies

Kim Hoskin
Forum|alt.badge.img+2
  • Moderator
  • June 10, 2026

Hi Sergi,
I recommend you to:
-Check if you have any Patch Update installed, if not please consider to installed the latest Patch Update on a test machine and check if that helps to avoid the issue.
-Raise a support case to seek further assistance for this issue, please include the details above, your region export, related files such as the XA switch sources/dlls used, so that my colleagues can look to recreate the behaviour. We may ask for additional traces such as CTF to be provided.

Regards,
Kim


  • Rocketeer
  • June 10, 2026

Hi Sergi , I would second what Kim said, Support line ticket would be best bet for this issue.