|
Problem IC87467
|
Status: Closed |
CLEAR NSE LOG TABLE ENTRIES ON THE VALUE SET FOR COMMITCOUNT |
| product: |
DB2 NET SEARCH / 5765F3800 / 970 - DB2 |
| Problem description: |
The index update crashed after processing 700,000 documents.This
lead to an abort of the complete index update, and the work of
indexing the previous 700,000 documents was lost. Then, the
index update is restarted with intermediate commits to at least
save the work until the crash. This didn't help to find the
'culprit' document as the log table remained unchanged and was
not updated to reflect the already processed documents. Even
though the index now contains several hundred thousand
successfully indexed documents, that work is completely redone. |
| Problem Summary: |
****************************************************************
* USERS AFFECTED: *
****************************************************************
* PROBLEM DESCRIPTION: *
* See Error Description *
****************************************************************
* RECOMMENDATION: *
* By default, NSE log table will be cleared only at the end of *
* update index processing. *
* If you would like to clear the log table once the commit is *
* finished ( for *
* incremental index update with commit count value set ), you *
* should set the environmental *
* variable "NSE_CLEAR_LOG_ON_COMMIT" before updating the *
* index. Here are the steps: *
* *
* export NSE_CLEAR_LOG_ON_COMMIT=TRUE *
* db2set DB2ENVLIST=NSE_CLEAR_LOG_ON_COMMIT *
* restart DB2/NSE *
* issue the incremental index update with COMMIT COUNT set *
**************************************************************** |
| Local Fix: |
1) Upgrade to ICC indexer and use maxbatchSize parameter to do
indexing in batches
2) Rely on any custom programming technique to find the document
causing the crash. |
| available fix packs: |
DB2 Version 9.7 Fix Pack 8 for Linux, UNIX, and Windows
DB2 Version 9.7 Fix Pack 9 for Linux, UNIX, and Windows
DB2 Version 9.7 Fix Pack 9a for Linux, UNIX, and Windows
DB2 Version 9.7 Fix Pack 10 for Linux, UNIX, and Windows
|
| Solution |
1) Upgrade to ICC indexer and use maxbatchSize parameter to do
indexing in batches
2) Rely on any custom programming technique to find the document
causing the crash.
3) Upgrade to V9.7 FP8 or later |
| Workaround |
not known / see Local fix |
| BUG-Tracking |
forerunner : APAR is sysrouted TO one or more of the following: IC95998 IC96141 IC96143 IC96144
follow-up : |
| Timestamps |
Date - problem reported :
Date - problem closed :
Date - last modified :
| 25.10.2012
28.03.2013
28.03.2013 |
| Problem solved at the following versions (IBM BugInfos) |
9.7.FP8
|
| Problem solved according to the fixlist(s) of the following version(s) |
| 9.7.0.8
|
 |
