Kbase P122224: Type I storage area does not appear to be reusing space
| Autor |
Progress Software Corporation - Progress |
| Acesso |
Público |
| Publicação |
5/1/2009 |
|
Status: Verified
SYMPTOM(s):
Type I storage area does not appear to be reusing space
Type I storage area seems to be using double the amount of blocks necessary to store records.
Type I storage area is growing significantly.
Dbanalys shows total # of records for that area and average size should require 4.9GB of disk space.
Prostrct statistics confirms the DB Engine is using 2 times the amount of disk space to store these records.
Chanalys on the storage area detects blocks that should not be on the RM Chain.
FACT(s) (Environment):
Area contains records only. No indexes stored in this area.
Very little fragmentation (1.0)
High scatter factor (4.4)
256 records per block for this storage area.
8K database block size.
Storage area is using default Create (150) and Toss (300) limits.
Progress version installed is 10.1A0201
OpenEdge 10.1A 64-bit
OpenEdge 10.1A02 Service Pack
IBM AIX
CAUSE:
Bug# OE00146288
CAUSE:
The problem was occurring due to a corrupt RM chain. The 2nd and 3rd blocks on the RM chain are corrupt (in that they should NOT be on the RM chain since they are not RM Blocks). Because of this, they will never use the remaining space on the RM chain as the code that searches the RM Chain looks no further when it encounters a non-RM block on the RM chain. Corruption was also detected in the Free chain of the index area associated with the table that had corruption on the RM chain.
FIX:
1. Dump the data from this area.
2. Truncate the area.
3. Then load the data back into to the area.
4. Then rebuild all the indexes in the index area to repair the free chain. (If the area containing the indexes for this table only contains indexes for the dumped table, then you could truncate this area as well).