Kbase P122152: Getting errors 8779 and 8780 performing idxfix when Database auditing is enabled
| Autor |
Progress Software Corporation - Progress |
| Acesso |
Público |
| Publicação |
23/06/2009 |
|
Status: Verified
SYMPTOM(s):
Getting errors 8779 and 8780 performing idxfix when Database auditing is enabled
SYSTEM ERROR: Index <index-num> (<owner-name>.<table-name>, <index-name>) contains <entry-count> entries, but <record-count> records found. (8779)
SYSTEM ERROR: Index <index-num> (<owner-name>.<table-name>, <index-name>): <n-pointers> index block pointers and <n-blocks> blocks found. (8780)
roll forward of AI files from database to hot standby database fails with error 1028
SYSTEM ERROR: Rollforward Apply AI note, BKUPDCTR=<val1>, note updctr=<val2>. (1028)
SYSTEM ERROR: Rollforward Apply AI note, BKUPDCTR=9998, note updctr=9998.
FACT(s) (Environment):
Event auditing policy enabled
A large number of users in the system having strong buffer pool contention
All Supported Operating Systems
OpenEdge 10.1A
OpenEdge 10.1B
CAUSE:
Bug # OE00135869
CAUSE:
This can be caused by a combination of three factors:
1) auditing is enabled
2) a large number of users in the system having strong buffer pool contention
3) event auditing policy enabled causing a premature buffer flush leading to index corruption.
(Event auditing in this case could be login/logout events, the use of database utilities, or a use of application-level auditing via ABL or SQL. )
This combination does not always lead to corruption but under some circumstances can cause index corruption.
FIX:
Upgrade to OpenEdge 10.1B01 or later.
Make sure to have a full index rebuild using proutil <databasename> -C idxbuild all to correct the index corruption. Idxfix online will not completely fix it.