Hi.
I'm currently investigating a problem at a site
and this is just a shoot-in-the-dark if someone
would have seen something similar before...
Environment :
DS20E 666 MHz
VMS 8.2 (with VMS82A_UPDATE V2.0)
Rdb V7.2-010 (7.2.0.1.0)
COBOL 2.7
A realy simple application crashes sometimes when
newly restared after the daily night-window when
all detached process are shutdown and restared.
The Cobol code calls a rather simple SELECT written
in Modular SQL.
Sometimes (well, most mornings), Rdb returnes -304
"SQLCODE_OUTOFRAN"
"Value is out of range for a host variable"
After one or two restarts of the detached process
all is fine and the error does not return until
the process have been restarted after the next
night-window again.
There are aprox 15 detached processes running the same
Cobol image, and the error comes from 1 up to a few of
these processes. The other runs just fine.
The Cobol code have been added with a dummy call
to the same MOD proc using hardcoded parameters
the points to a know existing record in the table
(it's a single table select), and the -304 error
commes always from *both* the dummy call *and* from
the real call, never from only one of them.
We are more or less lost at the moment.
We are currently planning to add 7.2.1.0.0 so
we ge the latest Rdb version, but there is nothing
in the rel.notes pointing to this problem.
If the upgrade of Rdb does not help, we plan to
try to build a reproducable case, if possible.
It feels as a very unpredictable error...
Well, as I said, just if someone have seen anything
similar... :-)
Best Regards
Jan-Erik Söderholm
Sweden.
|
