PJ40619: DASD CU FALL OVER TEST INPUT LIST SHUTDOW
Closed as program error.
See Problem Summary.
APAR NUMBER: PJ40619 PRODUCT: z/TPF FUNCTIONAL AREA: DASD SHIPPED IN PUT: 10 ABSTRACT: Due to PJ32026, unplanned module down ECBs were stalled waiting for I/O after a control unit failure. PACKAGE CONTENTS: Source Segments: (C) base/rt/cyen.asm Object Only Binaries: None. Configuration Independent Binaries: (C) base/obj/cyen.o (C) base/load/CYEN.so Support Files: base/lst/cyen.lst base/lst/CYEN.map OTHER BINARIES TO BUILD: NO None. COMMENTS: In a base-only environment, unplanned module down calls CPAD with force, quiesce, and cleanup options to perform all module down tasks in one call. If the module being taken offline is the last module for that LSS, the cleanup option performs I/O to update the subsystem status table (SSST) header and cache fast write ID (CFWID) table. This cleanup I/O is done prior to purging the module queue for the module going offline. This means other ECBs relying on this module to complete I/O must wait for the cleanup I/O to finish before the module down ECB can purge the module queue and IOBs can be processed. If multiple module down conditions occur at the same time, there is a scenario where module down ECBs can become stalled waiting for I/O to complete. For example, an unplanned module down ECB is processing module A, is performing the cleanup tasks, and some of the I/O to update the CFWID table or SSST header is queued to module B. Module B goes offline and the unplanned module down ECB for module B also has to perform cleanup tasks. Because of record locking, the cleanup for module B can't proceed until the cleanup for module A finishes. However, one or more IOBs for module A's cleanup tasks are stalled on module B and will never finish. This will lead to an input list shutdown condition and an IPL would be required to clear the condition.
SOLUTION: Segment cyen.asm is updated after label CYEN01B to perform module down processing in two steps. First, CPAD is called with the force and quiesce to force the module down and purge the module queue. After that is complete and the module is marked offline, CPAD is called a second time to perform the cleanup tasks including any I/O to update the SSST header and CFWID table. As a result of this change, module down ECBs should no longer be stalled waiting for I/O to complete. COREQS: NO None. MIGRATION CONSIDERATIONS: NO None. BUILD COMMANDS AND INSTRUCTIONS: YES #maketpf commands for linux maketpf -f CYEN cyen.o maketpf CYEN link TPF_VERIFY_LINK_REFS=NO maketpf CYEN link UPDATED INFORMATION UNITS: NO None. See your IBM representative if you need additional information. DOWNLOAD INSTRUCTIONS: http://www.ibm.com/software/htp/tpf/maint/maintztpf.html APAR URL: http://www.ibm.com/software/htp/tpf/ztpfmaint/put10/PJ40619.htm
Reported component name
Reported component ID
NoSpecatt / Xsystem
Last modified date
APAR is sysrouted FROM one or more of the following:
APAR is sysrouted TO one or more of the following:
Fixed component name
Fixed component ID
Applicable component levels
R110 PSY UP