by > Hello,
> I have been using CA Ingres 6.4/05 (sqs.ptx/00) on a Sequent system. Can
> anybody tell me if there is a known bug inserting into heap tables that
> have become extremely fragmented through inserts and deletes? The
> symptom is that the SQL says x rows have been inserted but when you
> select from the table the rows are not there.
> Thanks in advance.
There is a problem but it has nothing to do with fragmentation. In
the 6.4/05/02 Release Notes Supplement it is claimed to be corrected.
It is bug 64576.
"DML: Under certain circumstances (VAX cluster, or multiple
servers without shared cache), concurrent inserts from
different servers to a HEAP table with no secondary indexes
and default locking resulted in "orphan" pages in the middle
of the table, resulting in table corruption."
Interesting, we are seeing this with multiple servers and shared
cache. It seems to manifest itself in 2 forms. The first is
missing data which is what you have seen and we have seen once
the other is "E_DM9335_DM2F_END_FILE End of file was reached
during a read or write operation." which have no seen 3 times.
Our solution was to change the table structure.
--
Telephone: +61 3 344 6003 Fax +61 3 347 4803
Data Base Administrator, Information Technology Services (ASIS),
The University of Melbourne, Parkville, Australia. 3052.