DB2 - Problem description
Problem IC66117 | Status: Closed |
WHEN ROW LENGTH EXCEEDS TEMPORARY TABLESPACE PAGE SIZE, RETURN -901 ERROR INSTEAD OF MARKING DATABASE BAD | |
product: | |
DB2 FOR LUW / DB2FORLUW / 980 - DB2 | |
Problem description: | |
A database is marked bad when a query requires a row length that exceeds the temporary tablespace page size. The database then goes down and crash recovery is required. For example, when a query involving XML columns is calculated to have 12k row length, but the temporary tablespace page size is only 8k, the following error is seen: 2009-07-16-15.23.27.170841-240 I958361A1665 LEVEL: Severe PID : 1331306 TID : 12730 PROC : db2sysc 0 INSTANCE: dbinst1 NODE : 000 DB : SAMPLEDB APPHDL : 0-17122 APPID: ... AUTHID : SAMPLEID EDUID : 12730 EDUNAME: db2agent (SAMPLEDB) 0 FUNCTION: DB2 UDB, data management, setupDATInfo, probe:2860 RETCODE : ZRC=0x87040055=-2029780907=SQLD_PRGERR "Unknown PROGRAM ERROR" DIA8576C A data management services programming error occurred. MESSAGE : Descriptor row length exceeds maximum allowed. DATA #1 : unsigned integer, 4 bytes 12496 DATA #2 : unsigned integer, 4 bytes 8101 CALLSTCK: [0] 0x09000000062D6398 pdLogRC@glue15 + 0x118 [1] 0x0900000006EB1990 @101@setupDATInfo__FP8sqeAgentP8SQLD_TCBP12SQLD_TDIRRECP13SQLD_T DESCREC + 0x140 [2] 0x09000000059E6368 sqldCompleteTCBforDAT__FP8sqeAgentP8SQLD_TCBP16SQLB_OBJECT_DESCT 3iT5P12SQLD_TDIRRECP13SQLD_TDESCRECT5Ul + 0x28C [3] 0x09000000059E5B8C @101@sqldCreateDAT__FP8sqeAgentPUsN32UiT6UlT8iN310_P10SQLD_FIELD P12SQLD_COLINFOP15SQLD_RCTCOLINFOT10_PiPP8SQLD_TCBPcUsN221_T6 + 0x5C4 [4] 0x09000000059E4798 sqldInternalCreateTable__FP8sqeAgentP19SQLD_TABLECREATE_CB + 0x290 [5] 0x09000000059E43FC sqldTableCreateTemp__FP8sqeAgentPUsT2UcUlT5iP10SQLD_FIELDP12SQLD _COLINFOT7 + 0xF4 [6] 0x0900000007538B90 sqlri_mgjnAppendToBuffer__FP8sqlrr_cbP10sqlri_mgjn + 0x65C [7] 0x09000000057599DC sqlri_mj__FP8sqlrr_cb + 0x204 [8] 0x09000000058FBA40 sqlriSectInvoke__FP8sqlrr_cbP12sqlri_opparm + 0xFFFFFFFFFFFFF878 [9] 0x09000000059C12E8 sqlrr_process_fetch_request__FP14db2UCinterface + 0x13C The fix will ensure that instead of marking the database bad due to the severe error, the statement will fail with a -901 error, and the database will remain up. | |
Problem Summary: | |
**************************************************************** * USERS AFFECTED: * * All * **************************************************************** * PROBLEM DESCRIPTION: * * A database is marked bad when a query requires a row * * lengththat exceeds the temporary table space page size. * * Thedatabase then goes down nad crash recovery is required. * **************************************************************** * RECOMMENDATION: * * Upgrade to the latest fix pack. * **************************************************************** | |
Local Fix: | |
available fix packs: | |
DB2 Version 9.8 Fix Pack 3 for Linux, UNIX, and Windows | |
Solution | |
Problem is first fixed in DB2 UDB version 9.8 fix pack 3. | |
Workaround | |
not known / see Local fix | |
Timestamps | |
Date - problem reported : Date - problem closed : Date - last modified : | 08.02.2010 10.01.2011 10.01.2011 |
Problem solved at the following versions (IBM BugInfos) | |
9.8.FP3 | |
Problem solved according to the fixlist(s) of the following version(s) | |
9.8.0.3 |