DB2 - Problem description
Problem IC63280 | Status: Closed |
LOAD HANGS WITH ALL LOAD FORMATTERS(DB2LFRM) WAITING FOR SYNCUP OR TRANSFER BUFFER. | |
product: | |
DB2 FOR LUW / DB2FORLUW / 970 - DB2 | |
Problem description: | |
This problem with LOAD hanging is rarely hit, and it only affects LOAD using DEL files or ASC files (i.e. it does not affect other LOAD input types like IXF files, LOAD from CURSOR, etc.) The trigger is when customer's input data has very long records. This problem could be identified by checking the stack trace(call stack) of "load formatters (db2lfrmX)" and "load media reader (db2lmr)". "load media reader (db2lmr)" are stuck on stack trace like this: select sqlorqueInternal sqlorque2 sqluReadBufferFromQueue sqluMCContinueRestore sqluMCProcessRestoreStates sqluMCStartRestoreMediaController sqloCreateEDU ...... "load formatters (db2lfrmX)" are stuck on either stack trace: sqloWaitEDUWaitPost sqloSyncUp sqluRiddifyRecords sqluGetAndFormatRecords sqluldat sqloCreateEDU ...... or: msgrcv sqlorqueInternal sqlorque2 sqlulGetBuff sqlulcgf (for DEL file) or sqlulagr (for ASC file) sqlulcrr (for DEL file) or sqlularr (for ASC file) sqluGetAndFormatRecords sqluldat sqloCreateEDU ...... | |
Problem Summary: | |
**************************************************************** * USERS AFFECTED: * * All users of version 9.7 on Linux, Unix and Windows * * platforms. * **************************************************************** * PROBLEM DESCRIPTION: * * This problem is rarely hit and only affects LOAD using DEL * * files or ASC files(i.e. does not affect other LOAD input * * types like IXF files, LOAD from CURSOR etc.) * * The trigger is when customer's input data has very long * * records. * * * * This problem could be identified by checking the stack * * trace(call stack) of "load formatters(db2lfrmX)" and "load * * media reader(db2lmr)". * * * * "load media reader(db2lmr)" are stuck on stack trace like * * this: * * select * * sqlorqueInternal * * sqlorque2 * * sqluReadBufferFromQueue * * sqluMCContinueRestore * * sqluMCProcessRestoreStates * * sqluMCStartRestoreMediaController * * sqloCreateEDU * * ...... * * * * "load formatters(db2lfrmX)" are stuck on either stack trace: * * sqloWaitEDUWaitPost * * sqloSyncUp * * sqluRiddifyRecords * * sqluGetAndFormatRecords * * sqluldat * * sqloCreateEDU * * ...... * * or: * * msgrcv * * sqlorqueInternal * * sqlorque2 * * sqlulGetBuff * * sqlulcgf (for DEL file) or sqlulagr (for ASC file) * * sqlulcrr (for DEL file) or sqlularr (for ASC file) * * sqluGetAndFormatRecords * * sqluldat * * sqloCreateEDU * * ...... * **************************************************************** * RECOMMENDATION: * * Update to DB2 LUW Version 9.7 Fix Pack 6 or higher levels. * **************************************************************** | |
Local Fix: | |
1. Specify "CPU_PARALLELISM 1" in LOAD command; or 2. Specify "ANYORDER" option in LOAD command; or 3. Increase the amount of load buffer either explicitly via DATA BUFFER option or implicitly via increasing UTIL_HEAP_SZ database configuration parameter. (This will work around a specific problem instance, but it is possible to hit a different load hang with the new load buffer value.) | |
available fix packs: | |
DB2 Version 9.7 Fix Pack 6 for Linux, UNIX, and Windows | |
Solution | |
First fixed in DB2 LUW Version 9.7 Fix Pack 6. | |
Workaround | |
1. Specify "CPU_PARALLELISM 1" in LOAD command; or 2. Specify "ANYORDER" option in LOAD command; or 3. Increase the amount of load buffer either explicitly via DATA BUFFER option or implicitly via increasing UTIL_HEAP_SZ database configuration parameter.(This will work around a specific problem instance, but it is possible to hit a different load hang with the new load buffer value.) | |
BUG-Tracking | |
forerunner : APAR is sysrouted TO one or more of the following: IC84863 follow-up : | |
Timestamps | |
Date - problem reported : Date - problem closed : Date - last modified : | 20.09.2009 05.06.2012 05.06.2012 |
Problem solved at the following versions (IBM BugInfos) | |
9.7.FP6 | |
Problem solved according to the fixlist(s) of the following version(s) | |
9.7.0.6 |