AEIU: ILLOGIC not handled

What it means

CICS maps most VSAM problems onto named conditions. A missing key becomes NOTFND, a duplicate key becomes DUPREC, an I/O failure becomes IOERR. ILLOGIC is what is left over. It means VSAM reported an error that CICS has no specific condition for. For a file request, CICS puts the VSAM return code and error code into the EIBRCODE field of the EIB, so the details are in the EIB.

ILLOGIC also appears outside file control. The system programming browse commands, such as INQUIRE with START, NEXT and END, return it when the browse sequence is wrong. IBM APAR PI94527 describes the STAT transaction abending AEIU because it started a second browse of MQMONITOR resources while the first was still open.

The default action for ILLOGIC is to abend the task with AEIU. With RESP(WS-RESP) on the command, which implies NOHANDLE, the program gets control back and can test for DFHRESP(ILLOGIC). Older programs use HANDLE CONDITION ILLOGIC(paragraph).

Common causes

• An alternate index is out of step with its base cluster. IBM APAR DY47539, for CICS on z/VSE, describes AEIU with VSAM error code 90 and return code 8 when an alternate index entry pointed at a base record that no longer existed.
• The data set's structure is damaged. Updating a VSAM data set from a batch job while CICS has it open for update, under SHAREOPTIONS 3 or 4, is a classic way to break an index.
• An alternate index wasn't defined in the upgrade set, so changes to the base cluster never reached it.
• A system programming browse is out of sequence, such as a second INQUIRE START before END for the same resource type.

How to find the problem

1. The terminal message DFHAC2206 names the transaction. In the dump, EIBFN shows the command, EIBRSRCE the file name, and EIBRCODE the VSAM return code and error code.
2. Look up the VSAM return code and error code in the DFSMS macro instructions manual for your z/OS release. The error code usually points straight at the problem.
3. Run the transaction under CEDF. The EIB display shows EIBRCODE and EIBRESP2 for the failing command.
4. Check the cluster and any alternate index with IDCAMS EXAMINE (INDEXTEST and DATATEST), and use LISTCAT to confirm the alternate index is in the upgrade set.
5. Read the CICS job log for DFHFC messages about the same file. File control often reports the VSAM problem there too.

How to fix it

• Rebuild what is broken. Close the file in CICS, copy the base cluster to a new cluster with IDCAMS REPRO, rebuild each alternate index with BLDINDEX, then reopen the file.
• Stop batch jobs from updating the data set while CICS has it open for update. Close the file for the batch window, or move the data set to VSAM record level sharing.
• Define every alternate index with UPGRADE so it stays in step with the base cluster.
• Code RESP(WS-RESP) RESP2(WS-RESP2) on file commands and log EIBRCODE when the response is ILLOGIC, so the VSAM codes are captured without a dump.
• In programs that browse system resources, always issue the END browse, even when the first NEXT finds nothing.

What it looks like

DFHDU0203I 10/18/2026 11:27:03 CICSPRD1 A transaction dump was taken for dumpcode: AEIU, Dumpid: 1/0051.
DFHAC2206 11:27:03 CICSPRD1 Transaction CUS1 failed with abend AEIU. Updates to local recoverable resources backed out.

Last reviewed 2026-10-18 by the WND Software team. Written in our own words. For the full reference, check IBM’s CICS documentation for your release.