Kbase 17890: SYSTEM ERROR: <function>: Unable to build mandatory field ar
| Autor |
Progress Software Corporation - Progress |
| Acesso |
Público |
| Publicação |
10/05/1998 |
|
SYSTEM ERROR: <function>: Unable to build mandatory field ar
This error appears when we are in the process of building a mandatory
field array, which is a list of schema fields by position number and
dbkey. This list, or array is built on the stack, and an entry
is held for every field of the file. This process walks through all
the schema records, and assumes there is enough space on the stack
to hold a mandatory entry for every field of that file.
The "Unable to build mandatory field array. (1353)" appears when a
step in the process of building the mandatory list has gone wrong.
For example, if the stack size is not adequate to accommodate the list
of entries then we may see the error. Although we would expect
something like "increase -s", we don't always get that far when we
report some memory related errors and could throw this one out instead
Another potential cause is that one of the schema records we are
reading through is corrupted. Since we need to read through every
schema record on the mandatory list, it is plausible that one of them
could be corrupted.
As you can see, this process of reading through all the schema
records of a list created in memory leaves some potential areas of
concern, especially when reading through very large schema files.
If we encounter a memory munge (we get lost in memory) while walking
through this list, then we may see the error. If we clean up memory,
with a reboot for example, and we still see the error then we could
be blowing the memory stack, which is governed by the -s startup
parameter. Again, here we would expect a more specific error about the
stack size being to small, but we still may see the 1353.
Lastly, if it is not a munge in memory or a problem with the stack,
then we would have to lean towards a corrupt schema record, which
constitutes schema coruption. The only supported workaround for this
,if it does prove to be schema corruption, is restore a backup of the
database or dump and load.
SDA 5/4/98
Progress Software Technical Support Note # 17890