MainframeKit

ATNI

Task ended after a terminal errorCICS ABEND

CICS terminal error handling (DFHZNAC and the node error program) decided to end the task after an error on its terminal.

MOST LIKELY

The terminal session failed or was lost while the task was using it.

FIRST CHECK

Find the DFHZCxxxx message for the same terminal and time; it says what went wrong with the session.

TRY IT ▸

MainframeKit assistant

What it means

When a terminal session has an error, DFHZNAC picks actions for it and the node error program (DFHZNEP) can change them. If the action is to abend the task, the task attached to that terminal ends with ATNI, and a command it was running can get the TERMERR condition.

ATNI usually points to a problem with the terminal session rather than the program logic, though IBM has also reported it for a particular sequence of terminal commands. The DFHZCxxxx message written at the same time names the terminal and the error.

Typical causes

  • The terminal session failed or was lost while the task was using it.
  • The program issued a terminal command that was not allowed while an earlier abend, TERMERR or FREE was pending; IBM reported one case after SEND LAST WAIT followed by ISSUE DISCONNECT (message DFHZC3440).
  • The node error program was changed to request the task abend for this kind of error.

How to fix it

  1. Find the DFHZCxxxx message for the same terminal and time; it says what went wrong with the session.
  2. If the network or terminal dropped, the program is usually fine; check the session with the network team.
  3. If it follows a specific command sequence, check IBM support for APARs on your CICS release; for SEND LAST WAIT then ISSUE DISCONNECT, drop LAST or WAIT from the SEND.
  4. Review any changes to DFHZNEP that alter the abend-task action.
  5. Test RESP for TERMERR on terminal commands so the program can clean up.