DB2 - Problem description
Problem IC74968 | Status: Closed |
ABEND WHEN DROPPING SMALL TABLE OR REORGANIZING LOB OBJECT WHEN DB2_MEMORY_PROTECT ENABLED ON AIX WITH STORAGE KEY SUPPORT | |
product: | |
DB2 FOR LUW / DB2FORLUW / 970 - DB2 | |
Problem description: | |
The DB2_MEMORY_PROTECT registry variable is enabled, and DB2 is running on AIX with storage key support, which is one of the pre-conditions for DB2_MEMORY_PROTECT to take effect (please refer to Information Center for more detail). When these conditions are met, the DB2 instance may abend during one of the following actions. The common denominator is the presence of the sqlbPurgeOrFlushEMPs routine in the trap file. 1) Small Table Drop =================== This problem can only happen if the table to be dropped has dirty pages in one of the bufferpools defined on the current database, and the size of the table is relatively small compared to the size of the dirty lists for the bufferpool (the exact threshold is IBM Confidential and subject to change). In this case the table will qualify for fast dirty page purging, and the attempt to purge its dirty pages during DROP TABLE will cause an instance abend. The trap file will be similar to: EDU name : db2agent (SAMPLE) Signal #11 sqlpflog or sqlbGetPageLSN sqlpflog or sqlbCheckAndForceLogs sqlbPurgeOrFlushAllPagesInSmallRange sqlbPurgeOrFlushEMPs sqlbDMSDeleteObject 2) LOB Reorg ============ A similar trap may happen during a lob object reorganization. The trap file will be similar to: EDU name : db2agent (SAMPLE) Signal #11 sqlbAddToHateList_IUO sqlbPurgeAllPagesCommon sqlbPurgeOrFlushEMPs sqldReorgReplaceDMS sqldReorgCopyPhase Both problems only exist in DB2 UDB 9.7 FixPak 1 or newer. | |
Problem Summary: | |
**************************************************************** * USERS AFFECTED: * * ABEND WHEN DROPPING SMALL TABLE OR REORGANIZING LOB OBJECT * * WHEN DB2_MEMORY_PROTECT ENABLED ON AIX WITH STORAGE KEY * * SUPPORT * **************************************************************** * PROBLEM DESCRIPTION: * * The DB2_MEMORY_PROTECT registry variable is enabled, and DB2 * * is running on AIX with storage key support, which is one of * * the pre-conditions for DB2_MEMORY_PROTECT to take effect * * (please refer to Information Center for more detail). When * * these conditions are met, the DB2 instance may abend during * * one of the following actions. The common denominator is the * * presence of the sqlbPurgeOrFlushEMPs routine in the trap * * file. * * * * 1) Small Table Drop * * =================== * * This problem can only happen if the table to be dropped has * * dirty pages in one of the bufferpools defined on the current * * database, and the size of the table is relatively small * * compared to the size of the dirty lists for the bufferpool * * (the exact threshold is IBM Confidential and subject to * * change). In this case the table will qualify for fast dirty * * page purging, and the attempt to purge its dirty pages * * during DROP TABLE will cause an instance abend. The trap * * file will be similar to: * * * * EDU name : db2agent (SAMPLE) * * Signal #11 * * * * sqlpflog or sqlbGetPageLSN * * sqlpflog or sqlbCheckAndForceLogs * * sqlbPurgeOrFlushAllPagesInSmallRange * * sqlbPurgeOrFlushEMPs * * sqlbDMSDeleteObject * * * * 2) LOB Reorg * * ============ * * A similar trap may happen during a lob object * * reorganization. * * The trap file will be similar to: * * * * EDU name : db2agent (SAMPLE) * * Signal #11 * * * * sqlbAddToHateList_IUO * * sqlbPurgeAllPagesCommon * * sqlbPurgeOrFlushEMPs * * sqldReorgReplaceDMS * * sqldReorgCopyPhase * * * * Both problems only exist in DB2 UDB 9.7 FixPak 1 or newer. * **************************************************************** * RECOMMENDATION: * * Upgrade to DB2 UDB 9.7 FixPak 5. * **************************************************************** | |
Local Fix: | |
Disable DB2_MEMORY_PROTECT. | |
available fix packs: | |
DB2 Version 9.7 Fix Pack 5 for Linux, UNIX, and Windows | |
Solution | |
This APAR is first fixed in DB2 UDB 9.7 FixPak 5. | |
Workaround | |
not known / see Local fix | |
BUG-Tracking | |
forerunner : APAR is sysrouted TO one or more of the following: IC75711 follow-up : | |
Timestamps | |
Date - problem reported : Date - problem closed : Date - last modified : | 10.03.2011 23.12.2011 23.12.2011 |
Problem solved at the following versions (IBM BugInfos) | |
9.7. | |
Problem solved according to the fixlist(s) of the following version(s) | |
9.7.0.5 |