Kbase P141248: Hotspare database crashes during undo phase of crash recovery
| Autor |
Progress Software Corporation - Progress |
| Acesso |
Público |
| Publicação |
18/03/2009 |
|
Status: Verified
SYMPTOM(s):
OS copy of hotspare database crashes during undo phase of crash recovery with error 49
a restore of a probkup -norecover hotspare database crashes during undo phase of crash recovery with error 49
SYSTEM ERROR: Memory violation. (49)
Any subsequent copies of hotspare fail during crash recovery when accessed
After the hotspare copy is taken roll forward continues to run successfully against the hotspare database
If the hotspare database itself is accessed it also fails during the undo phase of crash recovery
FACT(s) (Environment):
OpenEdge 10.1x
OpenEdge 10.2A
All Supported Operating Systems
CAUSE:
Bug# OE00180748
CAUSE:
Roll Forward is miscalculating the size of the transaction table mask since OpenEdge 101a when JTA transactions were introduced. As a result, the variable overflowed and ended up with either a negative value or the "leftover bits" and the transaction table size and transaction mask were not set properly.
As a result, the transaction mask value gets set differently during roll forward than it was when the BI notes were created, depending on the size of the set of transactions needing logical backout, we would sometimes find we either couldn't allocate the table or the size was too small resulting in bi undo failure.
FIX:
This issue has been addressed in OpenEdge 10.1C04 and 10.2A01 Service Packs.
Workaround:
Re-create the hotspare target baseline. As with any DR plan, check the integrity of the backups on a regular basis.