Kbase P128363: Error 873 with negative updctr during ai roll forward.
| Autor |
Progress Software Corporation - Progress |
| Acesso |
Público |
| Publicação |
2/20/2008 |
|
Status: Unverified
FACT(s) (Environment):
All Supported Operating Systems
Progress 9.1E
OpenEdge 10.0x
OpenEdge 10.1A
SYMPTOM(s):
roll forward fails with error 873
SYSTEM ERROR: Physical redo, BKUPDCTR=<number>, note updctr=<number>. (873)
The number reported in the error message for "note updctr" is negative
Stack trace from _rfutil reads:
rlredo
dorollf
rlrollf
warmstrt
CAUSE:
Bug# OE00135882
CAUSE:
An overflow of the block update counter of the sequence block caused error 873, during the Redo phase of the warmstart database. This issue is particular to application environments that do millions of sequence operations between hotspare baselines.
FIX:
Upgrade to OpenEdge 10.1B01 or later where this issue has been addressed.
To recover:
There is nothing wrong with the source database. A new hotspare baseline needs to be taken.
In broad strokes, this entails:
0.) Ensure that you do not run out of ai files during this operation.
1.) Make a new full probkup of the production database, (can be online probkup)
2.) Restore it as a NEW hotspare database, (remove the previous instance first)
3.) Roll forward AI files that were created AFTER the (online) probkup was run.
Please refer to your company DR plan for futher information.